Add papers sheet from GitHub 5fca290

README.de.md

@@ -20,7 +20,7 @@
 
 <p align="center"><strong>Eine kuratierte Karte egozentrischer KI – die Datensätze, Benchmarks, Modelle und Werkzeuge hinter egozentrischem Sehen, verkörperter KI und Robotik, Video-Sprache, Langzeitgedächtnis, AR/VR und Hand-Objekt-Interaktion.</strong></p>
 
-<p align="center"><strong>456</strong> egozentrische Ressourcen — 125 Datensätze · 81 Benchmarks · 226 Modelle · 23 Toolkits</p>
+<p align="center"><strong>476</strong> egozentrische Ressourcen — 127 Datensätze · 90 Benchmarks · 235 Modelle · 23 Toolkits</p>
 
 ## Inhalt
 

README.es.md

@@ -20,7 +20,7 @@
 
 <p align="center"><strong>Un mapa curado de la IA egocéntrica: los conjuntos de datos, benchmarks, modelos y herramientas tras la visión egocéntrica, la IA encarnada y la robótica, el aprendizaje visión-lenguaje, la memoria de largo contexto, la RA/RV y la interacción mano-objeto.</strong></p>
 
-<p align="center"><strong>456</strong> recursos egocéntricos — 125 conjuntos de datos · 81 benchmarks · 226 modelos · 23 herramientas</p>
+<p align="center"><strong>476</strong> recursos egocéntricos — 127 conjuntos de datos · 90 benchmarks · 235 modelos · 23 herramientas</p>
 
 ## Qué incluye
 

README.fr.md

@@ -20,7 +20,7 @@
 
 <p align="center"><strong>Une carte sélective de l'IA égocentrique : les jeux de données, benchmarks, modèles et outils derrière la vision égocentrique, l'IA incarnée et la robotique, l'apprentissage vision-langage, la mémoire à long contexte, la RA/RV et l'interaction main-objet.</strong></p>
 
-<p align="center"><strong>456</strong> ressources égocentriques — 125 jeux de données · 81 benchmarks · 226 modèles · 23 outils</p>
+<p align="center"><strong>476</strong> ressources égocentriques — 127 jeux de données · 90 benchmarks · 235 modèles · 23 outils</p>
 
 ## Contenu
 

README.ja.md

@@ -20,7 +20,7 @@
 
 <p align="center"><strong>エゴセントリック AI の厳選マップ——エゴセントリック視覚、身体性 AI とロボティクス、ビデオ言語、長文脈記憶、AR/VR、手と物体の相互作用を支えるデータセット・ベンチマーク・モデル・ツールを収録。</strong></p>
 
-<p align="center"><strong>456</strong> エゴセントリック資源 — 125 データセット · 81 ベンチマーク · 226 モデル · 23 ツールキット</p>
+<p align="center"><strong>476</strong> エゴセントリック資源 — 127 データセット · 90 ベンチマーク · 235 モデル · 23 ツールキット</p>
 
 ## 収録内容
 

README.ko.md

@@ -20,7 +20,7 @@
 
 <p align="center"><strong>자기중심 AI의 엄선된 지도 — 자기중심 비전, 체화 AI와 로보틱스, 비디오-언어, 장문맥 기억, AR/VR, 손-물체 상호작용을 뒷받침하는 데이터셋·벤치마크·모델·도구를 담았습니다.</strong></p>
 
-<p align="center"><strong>456</strong> 자기중심 자원 — 125 데이터셋 · 81 벤치마크 · 226 모델 · 23 툴킷</p>
+<p align="center"><strong>476</strong> 자기중심 자원 — 127 데이터셋 · 90 벤치마크 · 235 모델 · 23 툴킷</p>
 
 ## 구성
 

README.md

@@ -25,6 +25,8 @@ configs:
   - config_name: catalog
     data_files: awesome-egocentric-atlas.csv
     default: true
+  - config_name: papers
+    data_files: awesome-egocentric-papers.csv
 ---
 
 ## Use this dataset
@@ -35,6 +37,9 @@ from datasets import load_dataset
 ds = load_dataset("cy0307/awesome-egocentric-atlas", split="train")
 print(len(ds), "resources")
 print(ds[0])
+
+papers = load_dataset("cy0307/awesome-egocentric-atlas", "papers", split="train")
+print(len(papers), "paper-linked resources")
 ```
 
 Each row is one catalogued resource. Columns:
@@ -56,7 +61,7 @@ Each row is one catalogued resource. Columns:
 | `tasks` | Task families (`; `-separated) |
 | `modalities` | Modalities (`; `-separated) |
 
-A richer nested-JSON version (with lanes and summary stats) is in [`site-data.json`](site-data.json). This dataset card and catalog are mirrored from the [GitHub repository](https://github.com/ChaoYue0307/awesome-egocentric-atlas) and the [interactive site](https://chaoyue0307.github.io/awesome-egocentric-atlas/); see the full README below for the curated tables and figures.
+The `papers` config is a paper-focused sheet containing only resources with a paper link. A richer nested-JSON version (with lanes and summary stats) is in [`site-data.json`](site-data.json). This dataset card and catalog are mirrored from the [GitHub repository](https://github.com/ChaoYue0307/awesome-egocentric-atlas) and the [interactive site](https://chaoyue0307.github.io/awesome-egocentric-atlas/); see the full README below for the curated tables and figures.
 
 ---
 <p align="center">
@@ -94,7 +99,7 @@ A richer nested-JSON version (with lanes and summary stats) is in [`site-data.js
   <a href="https://github.com/ChaoYue0307/awesome-egocentric-atlas/actions/workflows/validate.yml"><img alt="validate" src="https://github.com/ChaoYue0307/awesome-egocentric-atlas/actions/workflows/validate.yml/badge.svg"></a>
   <a href="https://chaoyue0307.github.io/awesome-egocentric-atlas/"><img alt="project site" src="https://img.shields.io/badge/site-GitHub%20Pages-067882"></a>
   <a href="https://huggingface.co/datasets/cy0307/awesome-egocentric-atlas"><img alt="Hugging Face mirror" src="https://img.shields.io/badge/Hugging%20Face-mirror-ffcc4d"></a>
-  <a href="data/resources.yml"><img alt="resources" src="https://img.shields.io/badge/resources-456-0097A7"></a>
+  <a href="data/resources.yml"><img alt="resources" src="https://img.shields.io/badge/resources-476-0097A7"></a>
   <a href="README.md#dataset-atlas"><img alt="datasets" src="https://img.shields.io/badge/datasets-vision%20%7C%20robotics%20%7C%20memory-344054"></a>
   <a href="README.md#models-tools-and-baselines"><img alt="models and tools" src="https://img.shields.io/badge/models-and%20tools-F5A623"></a>
   <a href="LICENSE"><img alt="license" src="https://img.shields.io/badge/license-MIT-667085"></a>
@@ -103,7 +108,7 @@ A richer nested-JSON version (with lanes and summary stats) is in [`site-data.js
 
 **Awesome Egocentric Atlas** maps the egocentric (first-person) AI landscape — datasets, benchmarks, models, and tools spanning egocentric vision, embodied AI and robotics, vision-language-action, world models, long-context memory, AR/VR, and hand-object interaction. Every entry shows its public-access status, so you can tell at a glance what you can download today and what is still just a paper.
 
-**Updated:** 2026-06-20.
+**Updated:** 2026-06-21.
 **Scope:** the main atlas is **human or animal first-person capture** from head, glasses, headset, body, wrist, handheld, or synchronized ego-exo rigs (where the ego view is central). Related but non-egocentric resources — robot-only datasets, multi-view robotic benchmarks, autonomous-driving 4D data, and general long-video reasoning — are listed separately under [Adjacent and Related Resources](#adjacent-and-related-resources) rather than in the main tables.
 
 <p align="center">
@@ -146,7 +151,7 @@ Prefer a browsable view? The [interactive site](https://chaoyue0307.github.io/aw
 
 | Signal | What it means for readers |
 | :--- | :--- |
-| 456 egocentric resources | 125 datasets, 81 benchmarks, 226 models, and 23 toolkits, plus a Project Aria collection hub — across vision, robotics, memory, and AR. Four related non-egocentric resources are listed separately. |
+| 476 egocentric resources | 127 datasets, 90 benchmarks, 235 models, and 23 toolkits, plus a Project Aria collection hub — across vision, robotics, memory, and AR. Three related non-egocentric resources are listed separately. |
 | 6 research lanes | Foundation video, procedure/action, hands and 3D, memory/reasoning, robotics/VLA, and AR/wearable sensing. |
 | 5 access states | `open`, `request`, `benchmark`, `partial`, and `watch` keep availability visible before you plan experiments. |
 | Machine-checked catalog | [`data/resources.yml`](data/resources.yml) is the source for type, year, status, URL, tasks, and provenance — and CI keeps the public artifacts in sync. |
@@ -262,6 +267,7 @@ Broad, large-scale first-person corpora that most egocentric pipelines pretrain
 | Resource | Released | Venue | Scale / signal | Best for | Status |
 | :--- | :---: | :---: | :--- | :--- | :---: |
 | [EgoCS-400K](https://arxiv.org/abs/2606.18180) | 2026-06 | arXiv | 400K+ first-person Counter-Strike gameplay videos (10K hours, 13 maps) with aligned actions, player state, camera motion, and game events | Action-conditioned interactive world models from first-person gameplay | watch |
+| [EgoInteract](https://arxiv.org/abs/2605.18214) | 2026-05 | arXiv | Controllable egocentric-video simulator and synthetic dataset with dense spatial and temporal annotations | Synthetic temporal segmentation, next-active-object detection, anticipation, and EHOI | watch |
 | [Ego-1K](https://huggingface.co/datasets/facebook/ego-1k) | 2026-03 | CVPR 2026 | Nearly 1,000 synchronized multiview egocentric videos from a custom 12-camera plus VR-headset rig | Dynamic 3D/4D scene understanding and novel view synthesis from ego rigs | open |
 | [EgoCrowds / CrowdEraser](https://arxiv.org/abs/2603.29036) | 2026-03 | arXiv | Semi-synthetic paired crowded/empty clips from real egocentric walking-tour video; CrowdEraser diffusion removes crowds for humanless walkthroughs | First-person walking-tour video editing and environment modeling | watch |
 | [Xperience-10M](https://huggingface.co/datasets/ropedia-ai/xperience-10m) | 2026-03 | Hugging Face | Ropedia release on Hugging Face; 10M experiences, 10K hours, six video streams, audio, stereo depth, camera pose, hand/body mocap, IMU, hierarchical language, ~1 PB total | Embodied AI, world models, robot learning from human experience, sensor fusion, 3D/4D understanding | request |
@@ -291,8 +297,13 @@ Egocentric human-manipulation and wrist-camera data aimed at vision-language-act
 | [EgoEngine](https://egoengine.github.io/) | 2026-06 | arXiv | Converts egocentric human manipulation videos into high-fidelity robot observation videos and executable robot action trajectories | Human-to-robot data generation, dexterous imitation | watch |
 | [EgoAERO](https://arxiv.org/abs/2606.08057) | 2026-06 | arXiv | Asset-free conversion from a single egocentric RGB-D demonstration; introduces EgoDex-R in the paper | Single-demo dexterous robot learning | watch |
 | [1M-HUGs / HUG](https://grasping.io/) | 2026-06 | arXiv | 1M frames / 27.8 hours of smart-glasses human grasps over 6,707 object instances, plus HUG-Bench with 90 unseen objects | Human grasp modeling and zero-shot robot grasping | open |
+| [HALOMI](https://arxiv.org/abs/2606.18772) | 2026-06 | arXiv | Extends UMI-style demonstration collection with egocentric head/wrist observations and head-hand trajectories for humanoid loco-manipulation | Active-perception humanoid manipulation from human demonstrations | watch |
+| [HumanoidArena](https://arxiv.org/abs/2606.17833) | 2026-06 | arXiv | Benchmark for egocentric hierarchical whole-body learning across seven leg-critical humanoid-object and humanoid-scene interaction tasks | Whole-body humanoid control from egocentric perception | watch |
+| [EgoSPT / SPOT](https://arxiv.org/abs/2605.20085) | 2026-05 | arXiv | Egocentric spatially prompted manipulation trajectories with first-frame object/target grounding and 3D end-effector motion | Spatially grounded manipulation trajectory prediction | watch |
 | [EgoVerse](https://egoverse.ai/) | 2026-04 | arXiv | 1,362 hours, 80K episodes, 1,965 tasks, 240 scenes, 2,087 demonstrators | Human demonstration scaling for robot learning and VLA | watch |
 | [EgoLive](https://arxiv.org/abs/2604.23570) | 2026-04 | arXiv | Large-scale real-world task-oriented egocentric routines for robot manipulation | Home service, retail, and real-world work-task manipulation | watch |
+| [GazeVLA](https://gazevla.github.io/) | 2026-04 | arXiv | VLA policy that pretrains on large-scale egocentric human data to capture gaze, intention, and action before robot fine-tuning | Gaze- and intent-aware robot manipulation | watch |
+| [WARPED](https://arxiv.org/abs/2604.10809) | 2026-04 | arXiv | Wrist-aligned rendering converts monocular egocentric human demonstrations into robot policy observations with 3D Gaussian Splatting | Cross-embodiment imitation from human ego video | watch |
 | [HRDexDB](https://arxiv.org/abs/2604.14944) | 2026-04 | arXiv | 1.4K human/robot grasping trials, tactile, multiview video, egocentric video streams | Cross-domain dexterous grasp learning | watch |
 | [AoE: Always-on Egocentric](https://arxiv.org/abs/2602.23893) | 2026-02 | arXiv | Always-on egocentric human-video collection pipeline and corpus for embodied AI | Scaling human-video data for robot learning | watch |
 | [EgoScale](https://arxiv.org/abs/2602.16710) | 2026-02 | arXiv | 20,854 hours of action-labeled egocentric human video with a human-to-robot two-stage transfer recipe and a log-linear data-scaling law | Scaling dexterous manipulation from human video | watch |
@@ -320,12 +331,16 @@ Fine-grained hand, object, contact, and 3D-pose datasets, including emerging eve
 | [Hand-4DGS](https://jeongminb.github.io/hand-4dgs/) | 2026-06 | arXiv | Feed-forward 4D hand reconstruction from egocentric video with 3D Gaussian Splatting, evaluated on H2O and ARCTIC | Fast egocentric 4D hand reconstruction | watch |
 | [A multimodal RGB/events FPV hand dataset](https://arxiv.org/abs/2606.10790) | 2026-06 | arXiv | Synthetic event-based first-person hand detection from EgoHands plus v2e | Event/RGB hand detection benchmarking | watch |
 | [DexGloveHOI](https://arxiv.org/abs/2605.21714) | 2026-05 | arXiv | 100K+ synchronized egocentric vision and on-glove IMU samples with marker-based mocap 3D hand-pose ground truth across dexterous daily manipulation | Vision-IMU fusion for 3D hand tracking | watch |
+| [EggHand](https://jyoun9.github.io/EggHand/) | 2026-05 | CVPR 2026 Findings | Multimodal foundation model for egocentric hand-pose forecasting over EgoExo4D-style video-language and action signals | Hand-pose forecasting and VLA action decoding | watch |
+| [Map-Mono-Ego](https://arxiv.org/abs/2605.20889) | 2026-05 | arXiv | Map-grounded global human-pose estimation from monocular egocentric video, with AIST-Living paired ego video and scanned environments | Scene-aware egocentric body-pose localization | watch |
 | [EgoEMG](https://arxiv.org/abs/2605.05712) | 2026-05 | arXiv | 41 participants, bilateral EMG, IMU, RGB, external RGB-D, mocap hand labels | EMG plus egocentric vision hand pose estimation | watch |
 | [EgoTouch / TouchAnything](https://jianyi2004.github.io/TouchAnything-Website/) | 2026-05 | arXiv | 302 tasks, 4,530 episodes with egocentric + dual wrist cameras, bimanual 3D hand pose, and dense tactile pressure maps | Tactile estimation and contact modeling from egocentric video | watch |
 | [TouchMoment](https://arxiv.org/abs/2604.12343) | 2026-04 | CVPR 2026 Findings | 4,021 egocentric videos, 8,456 annotated hand-object contact moments | Precise contact moment detection | watch |
 | [EgoFun3D](https://arxiv.org/abs/2604.11038) | 2026-04 | arXiv | 271 egocentric videos with 3D geometry, part segmentation, articulation and function-template annotations | Interactive 3D object modeling from ego video | watch |
+| [DP-DeGauss](https://arxiv.org/abs/2604.07986) | 2026-04 | arXiv | Dynamic probabilistic Gaussian decomposition for egocentric 4D scene reconstruction, separating background, hands, and objects | First-person 4D interaction reconstruction | watch |
 | [SHOW3D](https://arxiv.org/abs/2603.28760) | 2026-03 | CVPR 2026 | In-the-wild 3D hand-object interactions from a back-mounted multi-camera rig synced to a worn VR headset, with multi-view 3D shape/pose and text | In-the-wild 3D hand-object reconstruction | watch |
 | [EgoXtreme](https://arxiv.org/abs/2603.25135) | 2026-03 | CVPR 2026 | Smart-glasses egocentric 6D object-pose dataset across industrial, sports, and rescue scenes with extreme motion blur, dynamic lighting, and occlusion | Robust 6D object pose under extreme egocentric conditions | watch |
+| [AG-EgoPose](https://arxiv.org/abs/2603.25175) | 2026-03 | arXiv | Attention-guided egocentric 3D human-pose estimation from fisheye camera input with dual motion/spatial streams | Fisheye egocentric 3D pose estimation | watch |
 | [FEEL](https://arxiv.org/abs/2603.15847) | 2026-03 | arXiv | About 3M force-synchronized egocentric frames from kitchen manipulation with custom piezoresistive gloves | Contact-rich physical action and hand-object understanding | watch |
 | [Eva-3M / EvaPose](https://arxiv.org/abs/2602.23618) | 2026-02 | CVPR 2026 | 3.0M+ egocentric HPE frames, including 435K keypoint-visibility labels, plus visibility-aware pose estimation | Visibility-aware egocentric human pose | watch |
 | [WristPP](https://zhenqis123.github.io/WristPP/) | 2026-02 | CHI 2026 submission | Wrist-worn wide-FOV RGB system with a 133K-frame pose-pressure dataset from 20 subjects | Mobile hand pose and pressure interaction | watch |
@@ -366,12 +381,18 @@ Long-horizon daily-life capture and the question-answering benchmarks that probe
 | :--- | :---: | :---: | :--- | :--- | :---: |
 | [VL-MemKnG / WalkieKnowledgeT+](https://arxiv.org/abs/2606.17183) | 2026-06 | arXiv | Hybrid spatio-temporal knowledge graph plus segment memory for QA over long egocentric navigation trajectories | Navigation memory, long-horizon evidence retrieval, spatial QA | watch |
 | [V-RAGBench / CARVE](https://arxiv.org/abs/2606.13141) | 2026-06 | arXiv | Query, evidence-chunk, answer triplets for decoupled retrieval/generation evaluation in long egocentric video RAG | Retrieval-augmented long-video QA | watch |
+| [OVO-S-Bench](https://arxiv.org/abs/2606.03890) | 2026-06 | arXiv | 1,680 spatial-reasoning questions over 348 continuous egocentric videos with query timestamps and evidence intervals | Streaming spatial intelligence, tracking, simulation, and allocentric mapping | watch |
+| [VLESA](https://github.com/HanjiangHu/VLESA) | 2026-06 | arXiv | Goal-conditioned safety annotations over egocentric human-activity video for safety-aware assistance | Safety monitoring and situated activity assistance | watch |
 | [Causal-Plan-1M](https://arxiv.org/abs/2606.01810) | 2026-06 | arXiv | Reported million-scale corpus of explicit causal reasoning traces over egocentric videos | Causal and planning-oriented ego reasoning | watch |
 | [NoRA](https://arxiv.org/abs/2606.04806) | 2026-06 | arXiv | 1,420 first-person video clips for normative action reasoning with fact-reason-action support graphs | Grounded reasonableness and safety-oriented action generation | watch |
 | [Pause and Think](https://arxiv.org/abs/2606.00616) | 2026-06 | arXiv | Reasoning-centric training data and benchmark for video-grounded assistive action suggestions | Scene-grounded assistance, planning, and temporal consistency | watch |
 | [SuperMemory-VQA](https://arxiv.org/abs/2606.00825) | 2026-06 | arXiv | 52.9 hours of AI-glasses activity with RGB, audio, gaze, IMU, SLAM, and 4,853 human-verified QA pairs across object/location/intent/scene memory | Long-horizon memory for AR assistants | watch |
 | [EgoMemReason](https://arxiv.org/abs/2605.09874) | 2026-05 | arXiv | 500 questions over week-long egocentric video with entity, event, and behavior memory types | Memory-driven reasoning across sparse evidence over hours or days | watch |
 | [EgoBench](https://arxiv.org/abs/2605.27820) | 2026-05 | arXiv | 1,045 egocentric-video-grounded interactive tasks with tools and simulated users | Tool-using multimodal agents with dynamic interaction | watch |
+| [Minerva-Ego](https://github.com/google-deepmind/neptune) | 2026-05 | arXiv | Complex egocentric visual-reasoning benchmark with multi-step multimodal questions, dense reasoning traces, and spatiotemporal object masks | Grounded multi-step egocentric reasoning | open |
+| [EgoPro-Bench](https://arxiv.org/abs/2605.07299) | 2026-05 | arXiv | Personalized proactive-interaction benchmark with 2,400 evaluation videos, 12K+ training videos, and 12 domains | Streaming intent prediction and proactive assistance timing | watch |
+| [Pro2Assist](https://arxiv.org/abs/2605.04227) | 2026-05 | arXiv | Continuous step-aware proactive assistance with multimodal egocentric perception and an AR-glasses testbed | Real-time step-aware assistance | watch |
+| [EgoBabyVLM](https://arxiv.org/abs/2605.19130) | 2026-05 | arXiv | Benchmarking cross-modal learning from naturalistic infant and adult egocentric video, including Machine-DevBench and challenge tracks | Developmental egocentric VLM learning | watch |
 | [EgoIntrospect](https://ego-introspect.github.io/) | 2026-05 | arXiv | 180 hours from 60 subjects with synchronized video, audio, gaze, motion, physiological signals | Internal-state reasoning, affect, intent, cognitive memory for wearable assistants | watch |
 | [EgoExoMem](https://arxiv.org/abs/2605.18734) | 2026-05 | arXiv | 2.6K MCQs across synchronized ego-exo videos | Cross-view memory reasoning | watch |
 | [EgoCoT-Bench](https://dstardust.github.io/EgoCoT/) | 2026-05 | arXiv | 3,172 verifiable QA pairs over 351 videos with operation-centric rationale annotations | Grounded chain-of-thought and evidence consistency | open |
@@ -383,6 +404,7 @@ Long-horizon daily-life capture and the question-answering benchmarks that probe
 | [MA-EgoQA](https://ma-egoqa.github.io/) | 2026-03 | arXiv | 1.7K questions over multiple long-horizon egocentric streams | Multi-agent egocentric memory reasoning | open |
 | [Ego2Web](https://arxiv.org/abs/2603.22529) | 2026-03 | CVPR 2026 | Egocentric videos paired with web tasks requiring physical-scene understanding and online execution | Web agents grounded in first-person physical context | watch |
 | [Egocentric Co-Pilot](https://arxiv.org/abs/2603.01104) | 2026-03 | WWW 2026 | Web-native smart-glasses agent framework with temporal reasoning, context compression, speech/gaze intent, and streaming WebRTC/WebSocket pipelines | Assistive always-on egocentric web agents | watch |
+| [LifeEval](https://arxiv.org/abs/2603.00490) | 2026-03 | arXiv | 4,075 QA pairs over continuous first-person streams for real-time task-oriented human-AI collaboration in daily life | Daily-life multimodal assistance evaluation | watch |
 | [Gesture-Based Egocentric Video QA](https://arxiv.org/abs/2603.12533) | 2026-03 | CVPR 2026 | Egocentric video QA grounded in the camera wearer's pointing and deictic gestures | Gesture-grounded referential QA | watch |
 | [EgoIntent](https://arxiv.org/abs/2603.12147) | 2026-03 | arXiv | 3,014 steps over 15 daily-life scenarios for local intent (What), global intent (Why), and next-step plan (Next) | Step-level intent and anticipatory assistance | watch |
 | [SAW-Bench](https://arxiv.org/abs/2602.16682) | 2026-02 | arXiv | 786 Ray-Ban Meta smart-glasses videos and 2,071 QA pairs for observer-centric situated awareness | First-person situated spatial reasoning | watch |
@@ -425,8 +447,10 @@ Procedural and activity datasets, from modern industrial assembly to the classic
 | Resource | Released | Venue | Scale / signal | Best for | Status |
 | :--- | :---: | :---: | :--- | :--- | :---: |
 | [EgoMAGIC](https://arxiv.org/abs/2604.22036) | 2026-04 | arXiv | 3,355 egocentric field-medicine videos over 50 tasks from a head-mounted stereo camera with audio; 1.95M labels, 124 objects, action-detection challenge (Zenodo) | Field-medicine perception, action and object detection | open |
+| [PIE-V](https://arxiv.org/abs/2604.15134) | 2026-04 | arXiv | Mistake-aware procedural egocentric-video benchmark injecting plausible mistakes and recovery corrections across Ego-Exo4D scenarios | Procedural mistake detection and recovery reasoning | watch |
 | [IMPACT](https://arxiv.org/abs/2604.10409) | 2026-04 | arXiv | Ego-exo RGB-D industrial assembly dataset with bimanual, state, and anomaly annotations | Industrial assembly, procedural state tracking, anomaly recovery | watch |
 | [ENIGMA-360](https://arxiv.org/abs/2603.09741) | 2026-03 | arXiv | Ego-exo dataset for human behavior understanding in industrial scenarios with 360-degree and egocentric capture | Industrial ego-exo behavior understanding | watch |
+| [SAVA-X](https://arxiv.org/abs/2603.12764) | 2026-03 | CVPR 2026 | Ego-to-exo imitation-error detection over asynchronous, length-mismatched egocentric and exocentric videos, evaluated with EgoMe | Cross-view imitation-error detection | watch |
 | [PEDESTRIAN](https://arxiv.org/abs/2512.19190) | 2025-12 | arXiv | 340 first-person pavement videos covering 29 urban sidewalk obstacle types, with deep-learning detection baselines | Pedestrian-safety obstacle detection from first-person video | watch |
 | [IndEgo](https://huggingface.co/datasets/FraunhoferIPK/IndEgo) | 2025-11 | NeurIPS 2025 | ~197h egocentric (plus ~97h exocentric) industrial collaborative work over assembly, logistics, inspection, and repair; gaze, narration, sound, motion, hand pose, point clouds | Industrial egocentric assistants and procedure understanding | open |
 | [EgoEMS](https://arxiv.org/abs/2511.09894) | 2025-11 | AAAI 2026 | High-fidelity multimodal egocentric data for cognitive assistance in emergency medical services, capturing time-critical team actions | Real-time medical procedural assistance | watch |
@@ -495,8 +519,13 @@ Evaluation suites and label sets built on top of the raw datasets above.
 | [StreamMemBench](https://arxiv.org/abs/2606.14571) | 2026-06 | arXiv | EgoLife egocentric streams | Initial/follow-up task sequences testing evidence recall, feedback incorporation, and future assistance | watch |
 | [V-RAGBench / CARVE](https://arxiv.org/abs/2606.13141) | 2026-06 | arXiv | Long egocentric video chunks | Decoupled retrieval and generation evaluation for VideoRAG over evidence chunks | watch |
 | [VL-MemKnG / WalkieKnowledgeT+](https://arxiv.org/abs/2606.17183) | 2026-06 | arXiv | Long egocentric navigation trajectories | Temporally distributed spatial-memory QA with hybrid graph and segment retrieval | watch |
+| [OVO-S-Bench](https://arxiv.org/abs/2606.03890) | 2026-06 | arXiv | Continuous egocentric streams | Hierarchical spatial intelligence across perception, tracking, simulation, and allocentric mapping | watch |
+| [HumanoidArena](https://arxiv.org/abs/2606.17833) | 2026-06 | arXiv | Egocentric humanoid-control tasks | Hierarchical whole-body learning from egocentric perception | watch |
 | [Plan, Watch, Recover / EgoProactive](https://arxiv.org/abs/2606.04970) | 2026-06 | arXiv | EgoProactive plus Pro2Bench over five established benchmarks | Proactive procedural assistance, out-of-plan detection, and recovery guidance | watch |
 | [EgoBench](https://arxiv.org/abs/2605.27820) | 2026-05 | arXiv | Egocentric video tasks | Interactive multimodal tool-using agents | watch |
+| [Minerva-Ego](https://github.com/google-deepmind/neptune) | 2026-05 | arXiv | Egocentric videos with dense reasoning traces and object masks | Multi-step egocentric visual reasoning | open |
+| [EgoPro-Bench](https://arxiv.org/abs/2605.07299) | 2026-05 | arXiv | Personalized egocentric video streams | Proactive interaction, intent timing, and personalized assistance | watch |
+| [EgoBabyVLM](https://arxiv.org/abs/2605.19130) | 2026-05 | arXiv | Naturalistic infant and adult egocentric videos | Developmental cross-modal VLM evaluation | watch |
 | [Beyond Motion Primitives](https://arxiv.org/abs/2605.27464) | 2026-05 | arXiv | Ego4D | Head-mounted IMU benchmark for behavioral activity recognition on smart glasses | watch |
 | [Ego-METAS](https://maria-sanvil.github.io/Ego-METAS-website/) | 2026-05 | arXiv | EgoExo4D, CMU-MMAC, CaptainCook4D | Online multimodal, energy-aware temporal action segmentation across RGB, audio, gaze, IMU, and monochrome streams | watch |
 | [EgoProx](https://arxiv.org/abs/2605.24456) | 2026-05 | CVPR 2026 | Egocentric 3D proximity QA | Intention, exploration, exploitation, and chain-of-actions spatial reasoning for MLLMs | watch |
@@ -507,10 +536,13 @@ Evaluation suites and label sets built on top of the raw datasets above.
 | [EgoMemReason](https://arxiv.org/abs/2605.09874) | 2026-05 | arXiv | Week-long egocentric video | Entity, event, and behavior memory reasoning | watch |
 | [Ego2World](https://arxiv.org/abs/2605.13335) | 2026-05 | arXiv | HD-EPIC | Executable symbolic worlds from egocentric cooking video for belief-state planning | watch |
 | [EgoPoint-Bench](https://arxiv.org/abs/2604.21461) | 2026-04 | ACL 2026 | Simulated and real egocentric pointing samples | 11K+ QA items for referential reasoning and pointing-grounded object disambiguation | watch |
+| [PIE-V](https://arxiv.org/abs/2604.15134) | 2026-04 | arXiv | Ego-Exo4D-style procedural scenarios | Mistake-aware procedural-video evaluation with recovery corrections | watch |
 | [ReFocus / EM-QnF](https://nsubedi11.github.io/refocus) | 2026-04 | CVPR 2026 | Egocentric episodic-memory NLQ with user feedback | Interactive feedback refinement for ambiguous memory queries | watch |
 | [EgoEsportsQA](https://arxiv.org/abs/2604.12320) | 2026-04 | arXiv | First-person esports video | Fast virtual first-person perception and reasoning | watch |
 | [Audio Hallucination in Egocentric Video](https://arxiv.org/abs/2604.23860) | 2026-04 | ICASSP 2026 | Egocentric audio-visual video | 300 videos / 1,000 sound-focused questions probing audio hallucination in AV-LLMs | watch |
 | [EXPLORE-Bench](https://arxiv.org/abs/2603.09731) | 2026-03 | arXiv | Real first-person videos | Egocentric long-horizon scene-state prediction and reasoning | watch |
+| [LifeEval](https://arxiv.org/abs/2603.00490) | 2026-03 | arXiv | Continuous first-person streams | Real-time task-oriented human-AI collaboration in daily life | watch |
+| [SAVA-X](https://arxiv.org/abs/2603.12764) | 2026-03 | CVPR 2026 | EgoMe and asynchronous ego/exo video pairs | Ego-to-exo imitation-error detection | watch |
 | [MA-EgoQA](https://ma-egoqa.github.io/) | 2026-03 | arXiv | Multi-agent egocentric streams | Social, task coordination, theory-of-mind, temporal, environment QA | open |
 | [EgoAVU](https://github.com/facebookresearch/EgoAVU) | 2026-02 | CVPR 2026 | Egocentric audio-visual narrations | EgoAVU-Instruct (3M QAs) and EgoAVU-Bench (3K QAs) for audio-visual understanding (CVPR 2026 highlight) | open |
 | [SAW-Bench](https://arxiv.org/abs/2602.16682) | 2026-02 | arXiv | Ray-Ban Meta smart-glasses video | Observer-centric situated awareness and physically grounded spatial reasoning | watch |
@@ -550,9 +582,11 @@ Open models, baselines, and loaders you can build on directly.
 | :--- | :---: | :---: | :--- | :---: |
 | UNIEGO | 2026-06 | arXiv | Unified egocentric encoder distilled from nine teachers spanning ego-exo views, RGB, depth, skeleton, and foundation-model representations | [Paper](https://arxiv.org/abs/2606.20559) |
 | ActiveMimic | 2026-06 | arXiv | Egocentric human-video pretraining with active-perception signals for manipulation and VLA transfer | [Paper](https://arxiv.org/abs/2606.06194) |
+| VLESA | 2026-06 | arXiv | Goal-conditioned safety-assistance evaluation and baseline for monitoring human activities from egocentric video | [GitHub](https://github.com/HanjiangHu/VLESA) |
 | Continual Child-View Learning | 2026-06 | arXiv | Chronological multimodal learning from a child's egocentric video and speech stream | [Paper](https://arxiv.org/abs/2606.05115) |
 | Objects Before Words | 2026-06 | arXiv | Object-first language grounding from child-view egocentric video | [Paper](https://arxiv.org/abs/2606.12985) |
 | Watch Remember Reason | 2026-06 | arXiv | Human-view long-video understanding framework for MLLM watching, memory, and reasoning | [Paper](https://arxiv.org/abs/2606.07433) |
+| Pro2Assist | 2026-05 | arXiv | Continuous step-aware proactive assistance framework with multimodal egocentric perception and AR-glasses evaluation | [Paper](https://arxiv.org/abs/2605.04227) |
 | PhysBrain | 2025-12 | arXiv | Uses human egocentric data to bridge vision-language models toward physical intelligence and embodied control | [Paper](https://arxiv.org/abs/2512.16793) |
 | EgoM2P | 2025-06 | ICCV 2025 | Egocentric multimodal multitask pretraining over RGB, depth, gaze, and camera pose | [Paper](https://arxiv.org/abs/2506.07886) |
 | Exo2Ego / Ego-ExoClip | 2025-03 | AAAI 2026 | Transfers exocentric MLLM knowledge into egocentric video understanding with 1.1M synchronized ego-exo clip-text pairs and EgoIT instruction tuning | [Paper](https://arxiv.org/abs/2503.09143) |
@@ -578,6 +612,7 @@ Open models, baselines, and loaders you can build on directly.
 | MARS CASTLE | 2026-05 | CVPR 2026 EgoVis | Multimodal agentic reasoning with source selection for CASTLE challenge QA | [Paper](https://arxiv.org/abs/2605.18176) |
 | OSGNet + MLLM Reranking | 2026-05 | CVPR 2026 EgoVis | Champion Ego4D Episodic Memory Challenge solution for NLQ and GoalStep using MLLM reranking over OSGNet candidates | [GitHub](https://github.com/iLearn-Lab/CVPR25-OSGNet) |
 | OmniEgo-R2 | 2026-05 | CVPR 2026 EgoVis | Routed reasoning framework for EgoCross; second place in both Source-Limited and Open-Source tracks | [GitHub](https://github.com/Lee-zixu/OmniEgo-R2) |
+| Being-H0.7 | 2026-05 | arXiv | Latent world-action model from egocentric videos for future-aware reasoning and VLA policy learning | [Paper](https://arxiv.org/abs/2605.00078) |
 | Reflective Dialogue EgoCross | 2026-05 | CVPR 2026 EgoVis | Inference-time Teacher/Solver reflective dialogue for EgoCross support-set adaptation without fine-tuning | [Paper](https://arxiv.org/abs/2605.27885) |
 | EgoCross Domain-Wise Inference | 2026-05 | CVPR 2026 EgoVis | Nearly training-free source-limited inference strategy for EgoCross domain shift | [Paper](https://arxiv.org/abs/2606.00829) |
 | HD-EPIC Semantic-Visual Evidence | 2026-05 | CVPR 2026 EgoVis | HD-EPIC VQA challenge solution separating semantic and visual evidence | [Paper](https://arxiv.org/abs/2605.29402) |
@@ -656,6 +691,7 @@ Open models, baselines, and loaders you can build on directly.
 | Ego-Nav Co-training | 2026-06 | arXiv | Converts egocentric walking videos into robot-action datasets and co-trains a VLA with robot demos for mobile navigation | [Paper](https://arxiv.org/abs/2606.01951) |
 | EgoGuide | 2026-06 | arXiv | Synchronized wrist/head egocentric demonstration collection with online guidance and a gated egocentric residual policy | [Project](https://silicx.github.io/EgoGuide) |
 | Ego-Pi | 2026-06 | arXiv | VLA fine-tuning study across egocentric human and robot data using pi0.5 and dexterous five-finger embodiments | [Project](https://egopipaper.github.io/) |
+| HALOMI | 2026-06 | arXiv | Extends UMI-style egocentric collection for humanoid loco-manipulation with active perception from human demonstrations | [Paper](https://arxiv.org/abs/2606.18772) |
 | EgoTactile | 2026-06 | ICML 2026 spotlight | Egocentric video to full-hand grasp-pressure benchmark and diffusion/baseline models for everyday-object interactions | [Project](https://egotactile.github.io/) |
 | EgoPressDiff | 2026-06 | ICASSP 2026 | Conditional video diffusion for UV-domain egocentric hand-pressure maps using pose, mesh, and depth conditioning | [Project](https://egopressdiff.github.io/) |
 | Hand Trajectory Fusion for Ego NLQ | 2026-06 | CVPR 2026 EgoVis | Hand-trajectory encoder and cross-attention fusion for Ego4D Natural Language Query grounding | [Paper](https://arxiv.org/abs/2606.02962) |
@@ -673,17 +709,24 @@ Open models, baselines, and loaders you can build on directly.
 | EgoRelight | 2026-05 | arXiv | HMD-based egocentric human capture and illumination recovery for relightable avatars | [Paper](https://arxiv.org/abs/2605.28401) |
 | StableHand | 2026-05 | arXiv | Quality-aware flow-matching baseline for world-space dual-hand motion estimation from egocentric video | [Project](https://huajian-zeng.github.io/projects/stablehand/) |
 | HumanEgo | 2026-05 | arXiv | Zero-shot robot learning from minutes of human egocentric videos via entity-level hand-object representations and flow-matching policies | [Project](https://humanego-ai.github.io) |
+| EgoSPT / SPOT | 2026-05 | arXiv | Spatially prompted egocentric manipulation trajectory prediction from first-frame object/target grounding | [Paper](https://arxiv.org/abs/2605.20085) |
+| EggHand | 2026-05 | CVPR 2026 Findings | Multimodal egocentric hand-pose forecasting using video-language and VLA-style action decoding | [Project](https://jyoun9.github.io/EggHand/) |
+| Map-Mono-Ego | 2026-05 | arXiv | Map-grounded global human-pose estimation from monocular egocentric video and scanned environments | [Paper](https://arxiv.org/abs/2605.20889) |
 | EgoForce Hand Pose | 2026-05 | SIGGRAPH 2026 | Monocular egocentric 3D hand pose and shape reconstruction across fisheye, perspective, and wide-FOV camera models | [Project](https://dfki-av.github.io/EgoForce) |
 | EARL | 2026-05 | ICML 2026 | Analysis-guided RL framework for egocentric interaction reasoning and pixel grounding with coarse-to-fine parsing | [GitHub](https://github.com/yuggiehk/EARL) |
 | EgoExo-WM | 2026-05 | arXiv | Converts exocentric video into egocentric world-model training data using body-pose priors | [Project](https://vision.cs.utexas.edu/projects/EgoExo-WM/) |
 | MotionGRPO | 2026-05 | ICML 2026 | GRPO-based post-training for full-body 3D motion recovery from head-mounted device signals | [Paper](https://arxiv.org/abs/2605.05680) |
 | ActiveGlasses | 2026-04 | arXiv | Learns robot manipulation from smart-glasses ego-centric human demonstrations and transfers active vision to a robot perception arm | [Paper](https://arxiv.org/abs/2604.08534) |
+| GazeVLA | 2026-04 | arXiv | Pretrains on egocentric gaze, intention, and action signals before robot fine-tuning for manipulation | [Project](https://gazevla.github.io/) |
+| WARPED | 2026-04 | arXiv | Wrist-aligned rendering turns monocular egocentric human demonstrations into robot policy observations | [Paper](https://arxiv.org/abs/2604.10809) |
+| DP-DeGauss | 2026-04 | arXiv | Dynamic probabilistic Gaussian decomposition for egocentric 4D scene reconstruction of hands, objects, and background | [Paper](https://arxiv.org/abs/2604.07986) |
 | Personal Point of View 3DGS | 2026-04 | arXiv | Evaluation of dynamic 3D Gaussian splatting for egocentric scene reconstruction | [Paper](https://arxiv.org/abs/2604.23803) |
 | VGGT-Segmentor | 2026-04 | arXiv | Geometry-enhanced segmentation across egocentric and exocentric views | [Paper](https://arxiv.org/abs/2604.13596) |
 | Gaze-SoM HOI Anticipation | 2026-04 | arXiv | Gaze and set-of-mark prompting in VLLMs for hand-object-interaction anticipation from egocentric video | [Paper](https://arxiv.org/abs/2604.03667) |
 | EgoFlow | 2026-04 | arXiv | Gradient-guided flow matching for physically plausible 6DoF object-motion generation from egocentric video | [Paper](https://arxiv.org/abs/2604.01421) |
 | UniDex | 2026-03 | arXiv | Robot foundation suite for universal dexterous hand control learned from egocentric human videos | [Paper](https://arxiv.org/abs/2603.22264) |
 | PAWS | 2026-03 | arXiv | Articulation extraction from large-scale hand-object interactions in egocentric video | [Paper](https://arxiv.org/abs/2603.25539) |
+| AG-EgoPose | 2026-03 | arXiv | Attention-guided egocentric 3D human-pose estimation from fisheye camera input | [Paper](https://arxiv.org/abs/2603.25175) |
 | Static Scene Reconstruction from Dynamic Egocentric Videos | 2026-03 | arXiv | Mask-aware 3D reconstruction pipeline for long-form dynamic egocentric video | [Paper](https://arxiv.org/abs/2603.22450) |
 | EgoHOI World Model | 2026-03 | arXiv | Physics-informed egocentric world model that synthesizes contact-consistent hand-object interactions from action signals alone | [Paper](https://arxiv.org/abs/2603.13615) |
 | STAformer++ Affordance-Aware Anticipation | 2026-02 | arXiv | Integrates temporal attention, scene affordance memory, and interaction hotspots for short-term object-interaction anticipation on Ego4D and EPIC-KITCHENS | [Paper](https://arxiv.org/abs/2602.14837) |
@@ -808,7 +851,6 @@ These resources are **not first-person/egocentric**, but they are close neighbor
 | Resource | Released | Venue | Scale / signal | Why it is adjacent (not egocentric) | Status |
 | :--- | :---: | :---: | :--- | :--- | :---: |
 | [Seeing Across Views / MV-RoboBench](https://github.com/microsoft/MV-RoboBench) | 2025-10 | ICLR 2026 | 1.7K curated QA items over eight subtasks for multi-view spatial reasoning of VLMs in robotic manipulation (ICLR 2026) | Multi-camera robot scenes, not wearable capture | open |
-| [Minerva-Ego](https://github.com/google-deepmind/neptune) | 2025-05 | arXiv | Part of the Neptune / MINERVA long-video reasoning collection over web (YouTube) videos with manual reasoning traces | General long-video reasoning, not first-person capture | open |
 | [SEED4D](https://seed4d.github.io/) | 2024-12 | WACV 2025 | Synthetic ego-exo dynamic 4D generator and autonomous-driving dataset (16.8M images, vehicle cameras, LiDAR; WACV 2025) | Vehicle-egocentric driving data, not human/wearable | open |
 | [Open X-Embodiment / RT-X](https://robotics-transformer-x.github.io/) | 2023-10 | ICRA 2024 | 1M+ real robot trajectories, 22 robot embodiments, 60 pooled robot datasets, standardized RLDS; [unofficial Hugging Face mirror](https://huggingface.co/datasets/jxu124/OpenX-Embodiment) | Robot-mounted and wrist cameras, not human first-person; pairs with egocentric human-video VLA pretraining | open |
 

README.pt.md

@@ -20,7 +20,7 @@
 
 <p align="center"><strong>Um mapa curado da IA egocêntrica: os conjuntos de dados, benchmarks, modelos e ferramentas por trás da visão egocêntrica, da IA incorporada e da robótica, da aprendizagem visão-linguagem, da memória de longo contexto, da RA/RV e da interação mão-objeto.</strong></p>
 
-<p align="center"><strong>456</strong> recursos egocêntricos — 125 conjuntos de dados · 81 benchmarks · 226 modelos · 23 ferramentas</p>
+<p align="center"><strong>476</strong> recursos egocêntricos — 127 conjuntos de dados · 90 benchmarks · 235 modelos · 23 ferramentas</p>
 
 ## O que inclui
 

README.zh.md

@@ -20,7 +20,7 @@
 
 <p align="center"><strong>自我中心 AI 的精选地图——汇集自我中心视觉、具身智能与机器人、视频语言、长上下文记忆、AR/VR 以及手物交互背后的数据集、基准、模型与工具。</strong></p>
 
-<p align="center"><strong>456</strong> 自我中心资源 — 125 数据集 · 81 基准 · 226 模型 · 23 工具包</p>
+<p align="center"><strong>476</strong> 自我中心资源 — 127 数据集 · 90 基准 · 235 模型 · 23 工具包</p>
 
 ## 内容概览
 

app.js

@@ -11,6 +11,7 @@ const LANGS = [
 const I18N = {
   en: {
     skip: "Skip to catalog",
+    "language.label": "Language",
     "nav.catalog": "Catalog", "nav.lanes": "Lanes", "nav.access": "Access", "nav.readme": "README", "nav.milestones": "Milestones",
     "milestones.title": "Milestones", "milestones.desc": "The landmark works that shaped egocentric AI — from the field's origins to its current frontier.",
     "hero.lead": "A curated map of egocentric AI — the datasets, benchmarks, models, and tools behind egocentric vision, embodied AI and robotics, video-language, long-context memory, AR/VR, and hand-object interaction.",
@@ -32,11 +33,11 @@ const I18N = {
     "maintain.contributing": "Contributing guide", "maintain.schema": "Resource schema", "maintain.status": "Status policy", "maintain.workflow": "Maintenance workflow",
     "summary.total": "Total catalog", "summary.inscope": "In scope", "summary.adjacent": "Adjacent", "summary.open": "Open today", "summary.watch": "Watchlist", "summary.audit": "Last audit",
     "footer.tagline": "MIT licensed and free to use. Contributions welcome.",
-    "error.load": "Catalog failed to load", updated: "Updated {date}",
-    "bar.help": "Help translate", "bar.landing": "Landing page", "bar.hf": "Hugging Face mirror"
+    "error.load": "Catalog failed to load", updated: "Updated {date}"
   },
   zh: {
     skip: "跳到目录",
+    "language.label": "语言",
     "nav.catalog": "目录", "nav.lanes": "方向", "nav.access": "可获取性", "nav.readme": "README", "nav.milestones": "里程碑",
     "milestones.title": "里程碑", "milestones.desc": "塑造自我中心 AI 的里程碑工作——从领域起源到当前前沿。",
     "hero.lead": "自我中心 AI 的精选地图——汇集自我中心视觉、具身智能与机器人、视频语言、长上下文记忆、AR/VR 以及手物交互背后的数据集、基准、模型与工具。",
@@ -58,11 +59,11 @@ const I18N = {
     "maintain.contributing": "贡献指南", "maintain.schema": "资源结构", "maintain.status": "状态政策", "maintain.workflow": "维护流程",
     "summary.total": "目录总数", "summary.inscope": "范围内", "summary.adjacent": "相邻", "summary.open": "今日可用", "summary.watch": "关注列表", "summary.audit": "上次审核",
     "footer.tagline": "MIT 许可，免费使用，欢迎贡献。",
-    "error.load": "目录加载失败", updated: "更新于 {date}",
-    "bar.help": "帮助翻译", "bar.landing": "项目主页", "bar.hf": "Hugging Face 镜像"
+    "error.load": "目录加载失败", updated: "更新于 {date}"
   },
   es: {
     skip: "Saltar al catálogo",
+    "language.label": "Idioma",
     "nav.catalog": "Catálogo", "nav.lanes": "Áreas", "nav.access": "Acceso", "nav.readme": "README", "nav.milestones": "Hitos",
     "milestones.title": "Hitos", "milestones.desc": "Los trabajos de referencia que dieron forma a la IA egocéntrica, desde sus orígenes hasta la frontera actual.",
     "hero.lead": "Un mapa curado de la IA egocéntrica: los conjuntos de datos, benchmarks, modelos y herramientas tras la visión egocéntrica, la IA encarnada y la robótica, el aprendizaje visión-lenguaje, la memoria de largo contexto, la RA/RV y la interacción mano-objeto.",
@@ -84,11 +85,11 @@ const I18N = {
     "maintain.contributing": "Guía de contribución", "maintain.schema": "Esquema de recursos", "maintain.status": "Política de estados", "maintain.workflow": "Flujo de mantenimiento",
     "summary.total": "Catálogo total", "summary.inscope": "En alcance", "summary.adjacent": "Adyacentes", "summary.open": "Abiertos hoy", "summary.watch": "Lista de seguimiento", "summary.audit": "Última revisión",
     "footer.tagline": "Licencia MIT y de uso libre. Contribuciones bienvenidas.",
-    "error.load": "No se pudo cargar el catálogo", updated: "Actualizado {date}",
-    "bar.help": "Ayuda a traducir", "bar.landing": "Página principal", "bar.hf": "Espejo en Hugging Face"
+    "error.load": "No se pudo cargar el catálogo", updated: "Actualizado {date}"
   },
   fr: {
     skip: "Aller au catalogue",
+    "language.label": "Langue",
     "nav.catalog": "Catalogue", "nav.lanes": "Axes", "nav.access": "Accès", "nav.readme": "README", "nav.milestones": "Jalons",
     "milestones.title": "Jalons", "milestones.desc": "Les travaux marquants qui ont façonné l'IA égocentrique, des origines du domaine à sa frontière actuelle.",
     "hero.lead": "Une carte sélective de l'IA égocentrique : les jeux de données, benchmarks, modèles et outils derrière la vision égocentrique, l'IA incarnée et la robotique, l'apprentissage vision-langage, la mémoire à long contexte, la RA/RV et l'interaction main-objet.",
@@ -110,11 +111,11 @@ const I18N = {
     "maintain.contributing": "Guide de contribution", "maintain.schema": "Schéma des ressources", "maintain.status": "Politique des statuts", "maintain.workflow": "Flux de maintenance",
     "summary.total": "Catalogue total", "summary.inscope": "Dans le périmètre", "summary.adjacent": "Adjacentes", "summary.open": "Ouvertes aujourd'hui", "summary.watch": "Liste de veille", "summary.audit": "Dernière révision",
     "footer.tagline": "Licence MIT, libre d'utilisation. Contributions bienvenues.",
-    "error.load": "Échec du chargement du catalogue", updated: "Mis à jour le {date}",
-    "bar.help": "Aider à traduire", "bar.landing": "Page d'accueil", "bar.hf": "Miroir Hugging Face"
+    "error.load": "Échec du chargement du catalogue", updated: "Mis à jour le {date}"
   },
   de: {
     skip: "Zum Katalog springen",
+    "language.label": "Sprache",
     "nav.catalog": "Katalog", "nav.lanes": "Bereiche", "nav.access": "Zugang", "nav.readme": "README", "nav.milestones": "Meilensteine",
     "milestones.title": "Meilensteine", "milestones.desc": "Die wegweisenden Arbeiten, die egozentrische KI geprägt haben – von den Ursprüngen bis zur aktuellen Front.",
     "hero.lead": "Eine kuratierte Karte egozentrischer KI – die Datensätze, Benchmarks, Modelle und Werkzeuge hinter egozentrischem Sehen, verkörperter KI und Robotik, Video-Sprache, Langzeitgedächtnis, AR/VR und Hand-Objekt-Interaktion.",
@@ -136,11 +137,11 @@ const I18N = {
     "maintain.contributing": "Beitragsleitfaden", "maintain.schema": "Ressourcenschema", "maintain.status": "Status-Richtlinie", "maintain.workflow": "Wartungsablauf",
     "summary.total": "Gesamtkatalog", "summary.inscope": "Im Fokus", "summary.adjacent": "Angrenzend", "summary.open": "Heute offen", "summary.watch": "Beobachtungsliste", "summary.audit": "Letzte Prüfung",
     "footer.tagline": "MIT-Lizenz, frei nutzbar. Beiträge willkommen.",
-    "error.load": "Katalog konnte nicht geladen werden", updated: "Aktualisiert am {date}",
-    "bar.help": "Beim Übersetzen helfen", "bar.landing": "Startseite", "bar.hf": "Hugging-Face-Spiegel"
+    "error.load": "Katalog konnte nicht geladen werden", updated: "Aktualisiert am {date}"
   },
   ja: {
     skip: "カタログへスキップ",
+    "language.label": "言語",
     "nav.catalog": "カタログ", "nav.lanes": "研究分野", "nav.access": "アクセス", "nav.readme": "README", "nav.milestones": "マイルストーン",
     "milestones.title": "マイルストーン", "milestones.desc": "エゴセントリック AI を形づくった画期的な研究——分野の起源から最前線まで。",
     "hero.lead": "エゴセントリック AI の厳選マップ——エゴセントリック視覚、身体性 AI とロボティクス、ビデオ言語、長文脈記憶、AR/VR、手と物体の相互作用を支えるデータセット・ベンチマーク・モデル・ツールを収録。",
@@ -162,11 +163,11 @@ const I18N = {
     "maintain.contributing": "貢献ガイド", "maintain.schema": "資源スキーマ", "maintain.status": "状態ポリシー", "maintain.workflow": "メンテナンス手順",
     "summary.total": "カタログ総数", "summary.inscope": "対象内", "summary.adjacent": "隣接", "summary.open": "本日公開", "summary.watch": "ウォッチリスト", "summary.audit": "最終確認",
     "footer.tagline": "MIT ライセンス、自由に利用可能。貢献歓迎。",
-    "error.load": "カタログの読み込みに失敗しました", updated: "更新日 {date}",
-    "bar.help": "翻訳に協力", "bar.landing": "ランディングページ", "bar.hf": "Hugging Face ミラー"
+    "error.load": "カタログの読み込みに失敗しました", updated: "更新日 {date}"
   },
   ko: {
     skip: "카탈로그로 건너뛰기",
+    "language.label": "언어",
     "nav.catalog": "카탈로그", "nav.lanes": "연구 분야", "nav.access": "접근성", "nav.readme": "README", "nav.milestones": "이정표",
     "milestones.title": "이정표", "milestones.desc": "자기중심 AI를 형성한 획기적 연구 — 분야의 기원부터 현재 최전선까지.",
     "hero.lead": "자기중심 AI의 엄선된 지도 — 자기중심 비전, 체화 AI와 로보틱스, 비디오-언어, 장문맥 기억, AR/VR, 손-물체 상호작용을 뒷받침하는 데이터셋·벤치마크·모델·도구를 담았습니다.",
@@ -188,11 +189,11 @@ const I18N = {
     "maintain.contributing": "기여 가이드", "maintain.schema": "자원 스키마", "maintain.status": "상태 정책", "maintain.workflow": "유지보수 절차",
     "summary.total": "전체 카탈로그", "summary.inscope": "범위 내", "summary.adjacent": "인접", "summary.open": "오늘 공개", "summary.watch": "관심 목록", "summary.audit": "최근 점검",
     "footer.tagline": "MIT 라이선스, 자유롭게 사용하세요. 기여 환영.",
-    "error.load": "카탈로그를 불러오지 못했습니다", updated: "업데이트 {date}",
-    "bar.help": "번역 돕기", "bar.landing": "랜딩 페이지", "bar.hf": "Hugging Face 미러"
+    "error.load": "카탈로그를 불러오지 못했습니다", updated: "업데이트 {date}"
   },
   pt: {
     skip: "Ir para o catálogo",
+    "language.label": "Idioma",
     "nav.catalog": "Catálogo", "nav.lanes": "Áreas", "nav.access": "Acesso", "nav.readme": "README", "nav.milestones": "Marcos",
     "milestones.title": "Marcos", "milestones.desc": "Os trabalhos marcantes que moldaram a IA egocêntrica — das origens do campo à sua fronteira atual.",
     "hero.lead": "Um mapa curado da IA egocêntrica: os conjuntos de dados, benchmarks, modelos e ferramentas por trás da visão egocêntrica, da IA incorporada e da robótica, da aprendizagem visão-linguagem, da memória de longo contexto, da RA/RV e da interação mão-objeto.",
@@ -214,8 +215,7 @@ const I18N = {
     "maintain.contributing": "Guia de contribuição", "maintain.schema": "Esquema de recursos", "maintain.status": "Política de estados", "maintain.workflow": "Fluxo de manutenção",
     "summary.total": "Catálogo total", "summary.inscope": "No escopo", "summary.adjacent": "Adjacentes", "summary.open": "Abertos hoje", "summary.watch": "Lista de observação", "summary.audit": "Última auditoria",
     "footer.tagline": "Licença MIT e de uso livre. Contribuições bem-vindas.",
-    "error.load": "Falha ao carregar o catálogo", updated: "Atualizado em {date}",
-    "bar.help": "Ajude a traduzir", "bar.landing": "Página inicial", "bar.hf": "Espelho Hugging Face"
+    "error.load": "Falha ao carregar o catálogo", updated: "Atualizado em {date}"
   }
 };
 
@@ -299,8 +299,10 @@ const els = {
   lanes: document.querySelector("#lane-grid"),
   statuses: document.querySelector("#status-list"),
   empty: document.querySelector("#empty-state"),
-  langBar: document.querySelector("#lang-bar"),
-  milestoneLinks: document.querySelector("#milestone-link-layer")
+  languageSwitcher: document.querySelector("#language-switcher"),
+  languageMenu: document.querySelector("#language-menu"),
+  languageCurrent: document.querySelector("#language-current"),
+  milestoneBoard: document.querySelector("#milestone-board")
 };
 
 function titleize(value) {
@@ -341,36 +343,62 @@ function applyStaticI18n() {
   document.documentElement.lang = state.lang;
 }
 
-function buildLangBar() {
-  if (!els.langBar) return;
-  els.langBar.replaceChildren();
-  const frag = document.createDocumentFragment();
-  LANGS.forEach(([code, label], index) => {
-    if (index > 0) frag.appendChild(document.createTextNode(" · "));
-    const link = document.createElement("a");
-    link.href = "#";
-    link.textContent = label;
-    link.setAttribute("lang", code);
-    if (code === state.lang) link.className = "active";
-    link.addEventListener("click", (event) => {
-      event.preventDefault();
-      setLang(code);
+function langHref(code) {
+  const p = new URLSearchParams(location.search);
+  if (code === "en") {
+    p.delete("lang");
+  } else {
+    p.set("lang", code);
+  }
+  const qs = p.toString();
+  return `${location.pathname}${qs ? `?${qs}` : ""}${location.hash}`;
+}
+
+function createLanguageLink(code, label) {
+  const link = document.createElement("a");
+  link.href = langHref(code);
+  link.textContent = label;
+  link.setAttribute("lang", code);
+  if (code === state.lang) {
+    link.classList.add("active");
+    link.setAttribute("aria-current", "true");
+  }
+  link.addEventListener("click", (event) => {
+    event.preventDefault();
+    setLang(code);
+    if (els.languageSwitcher) els.languageSwitcher.open = false;
+  });
+  return link;
+}
+
+function buildLanguageSwitcher() {
+  const activeLang = LANGS.find(([code]) => code === state.lang) || LANGS[0];
+
+  if (els.languageCurrent) {
+    els.languageCurrent.textContent = activeLang[1];
+    els.languageCurrent.setAttribute("lang", activeLang[0]);
+  }
+
+  if (els.languageMenu) {
+    els.languageMenu.replaceChildren();
+    LANGS.forEach(([code, label]) => {
+      els.languageMenu.appendChild(createLanguageLink(code, label));
     });
-    frag.appendChild(link);
+  }
+}
+
+function bindLanguageSwitcher() {
+  if (!els.languageSwitcher) return;
+  document.addEventListener("click", (event) => {
+    if (!els.languageSwitcher.contains(event.target)) {
+      els.languageSwitcher.open = false;
+    }
   });
-  const extra = [
-    [t("bar.help"), "https://github.com/ChaoYue0307/awesome-egocentric-atlas/blob/main/CONTRIBUTING.md"],
-    [t("bar.landing"), "https://github.com/ChaoYue0307/awesome-egocentric-atlas"],
-    [t("bar.hf"), "https://huggingface.co/datasets/cy0307/awesome-egocentric-atlas"]
-  ];
-  extra.forEach(([label, href]) => {
-    frag.appendChild(document.createTextNode(" · "));
-    const link = document.createElement("a");
-    link.href = href;
-    link.textContent = label;
-    frag.appendChild(link);
+  document.addEventListener("keydown", (event) => {
+    if (event.key === "Escape") {
+      els.languageSwitcher.open = false;
+    }
   });
-  els.langBar.appendChild(frag);
 }
 
 function setLang(lang) {
@@ -379,7 +407,7 @@ function setLang(lang) {
   localStorage.setItem("aea-lang", lang);
   syncUrl();
   applyStaticI18n();
-  buildLangBar();
+  buildLanguageSwitcher();
   if (state.data) {
     renderStats();
     renderSummary();
@@ -538,26 +566,64 @@ function renderStatuses() {
   });
 }
 
-function renderMilestoneLinks() {
-  if (!els.milestoneLinks || !state.data.milestone_layout) return;
-  const layout = state.data.milestone_layout;
-  const width = Number(layout.width) || 1280;
-  const height = Number(layout.height) || 2322;
-  els.milestoneLinks.replaceChildren();
-
-  (layout.cells || []).forEach((cell) => {
-    const link = document.createElement("a");
-    link.className = "milestone-cell-link";
-    link.href = cell.url;
-    link.target = "_blank";
-    link.rel = "noopener noreferrer";
-    link.title = `${cell.name} (${cell.date})`;
-    link.setAttribute("aria-label", `${cell.name} (${cell.date})`);
-    link.style.left = `${(Number(cell.x) / width) * 100}%`;
-    link.style.top = `${(Number(cell.y) / height) * 100}%`;
-    link.style.width = `${(Number(cell.width) / width) * 100}%`;
-    link.style.height = `${(Number(cell.height) / height) * 100}%`;
-    els.milestoneLinks.appendChild(link);
+function milestoneEra(item) {
+  const year = Number(String(item.date || "").slice(0, 4));
+  if (year <= 2015) return { id: "origins", range: "2009-2015", title: "Origins", copy: "Daily-life activity, gaze, and hands become measurable egocentric signals." };
+  if (year <= 2022) return { id: "scale", range: "2020-2022", title: "Modern Scale", copy: "Large benchmarks, smart-glasses sensing, geometry, and video-language pretraining mature." };
+  if (year <= 2024) return { id: "reasoning", range: "2023-2024", title: "Reasoning & Robotics", copy: "Long-form reasoning, ego-exo capture, AR hand-object tracking, and robot interfaces converge." };
+  if (year === 2025) return { id: "daily", range: "2025", title: "Daily Life to VLA", copy: "Personal memory and egocentric demonstrations begin feeding robot policies." };
+  return { id: "frontier", range: "2026", title: "World-Model Frontier", copy: "Egocentric corpora, world models, and scaling laws push embodied AI outward." };
+}
+
+function renderMilestones() {
+  if (!els.milestoneBoard || !state.data.milestones) return;
+  const eras = [];
+  const eraMap = new Map();
+
+  state.data.milestones.forEach((item) => {
+    const era = milestoneEra(item);
+    if (!eraMap.has(era.id)) {
+      eraMap.set(era.id, { ...era, items: [] });
+      eras.push(eraMap.get(era.id));
+    }
+    eraMap.get(era.id).items.push(item);
+  });
+
+  els.milestoneBoard.replaceChildren();
+  eras.forEach((era, index) => {
+    const section = document.createElement("section");
+    section.className = "milestone-era";
+    section.setAttribute("aria-label", `${era.range} ${era.title}`);
+    const cards = era.items.map((item) => {
+      const kind = titleize(item.kind);
+      const origin = item.origin ? `From ${item.origin}.` : "";
+      const label = `${item.name}, ${item.date}, ${kind}. ${origin} ${item.note || ""}`.trim();
+      return `
+        <a class="milestone-card" href="${escapeHtml(item.url)}" target="_blank" rel="noopener noreferrer" title="${escapeHtml(label)}" aria-label="${escapeHtml(label)}">
+          <span class="milestone-card-media">
+            <img src="${escapeHtml(item.image || "assets/awesome-egocentric-logo.png")}" loading="lazy" decoding="async" alt="">
+          </span>
+          <span class="milestone-card-meta">
+            <span class="chip milestone-card-date">${escapeHtml(item.date)}</span>
+            <span class="chip milestone-card-kind">${escapeHtml(kind)}</span>
+          </span>
+          <strong class="milestone-card-title">${escapeHtml(item.name)}</strong>
+          ${item.origin ? `<span class="milestone-card-origin">${escapeHtml(item.origin)}</span>` : ""}
+          ${item.note ? `<span class="milestone-card-note">${escapeHtml(item.note)}</span>` : ""}
+        </a>
+      `;
+    }).join("");
+
+    section.innerHTML = `
+      <div class="milestone-era-head">
+        <p class="milestone-era-kicker">Era ${index + 1}</p>
+        <span class="milestone-era-range">${escapeHtml(era.range)}</span>
+        <span class="milestone-era-title">${escapeHtml(era.title)}</span>
+        <p class="milestone-era-copy">${escapeHtml(era.copy)}</p>
+      </div>
+      <div class="milestone-cards">${cards}</div>
+    `;
+    els.milestoneBoard.appendChild(section);
   });
 }
 
@@ -584,7 +650,8 @@ function bindFilters() {
 
 async function init() {
   applyStaticI18n();
-  buildLangBar();
+  buildLanguageSwitcher();
+  bindLanguageSwitcher();
   const response = await fetch("./site-data.json");
   state.data = await response.json();
   renderStats();
@@ -592,7 +659,7 @@ async function init() {
   renderFilters();
   renderLanes();
   renderStatuses();
-  renderMilestoneLinks();
+  renderMilestones();
   applyFiltersToForm();
   bindFilters();
   renderRows();

assets/README.md

@@ -34,7 +34,7 @@ The README and GitHub Pages site embed high-resolution PNG exports of the SVG fi
 - `awesome-egocentric-atlas-map.png` — `3840 x 2190`
 - `awesome-egocentric-reader-route.png` — `3840 x 1380`
 - `awesome-egocentric-timeline.png` — `3840 x 1680`
-- `awesome-egocentric-milestones.png` — `3840 x 6966`
+- `awesome-egocentric-milestones.png` — `4800 x 8907`
 - `awesome-egocentric-task-matrix.png` — `3840 x 2220`
 - `awesome-egocentric-access-funnel.png` — `3840 x 1560`
 
@@ -52,7 +52,7 @@ Visual system (shared across the figures):
 
 - `awesome-egocentric-timeline.svg` — resources grouped into five eras (`<=2018`, `2019-2021`, `2022-2023`, `2024-2025`, `2026 (H1)`), drawn as a bar chart with a trend line through the bar tops to make the rise explicit. Bar heights come from the egocentric `year` counts in `data/resources.yml`; the latest era is amber and counts only the first half of 2026 (January to June), labelled accordingly.
 - `awesome-egocentric-access-funnel.svg` — the egocentric resources by `status` (`open`, `watch`, `partial`, `benchmark`, `request`) as a tapering funnel, so readers see how much is usable today versus still unverified. Bar widths and counts come from the catalog.
-- `awesome-egocentric-milestones.svg` — a work-specific milestone poster generated from the `milestone`, `milestone_note`, and `milestone_image` fields in `data/resources.yml`. The factual labels, dates, and categories are vector text, while the visual panels come from the local `milestones/*.png` images and are rendered uncropped.
+- `awesome-egocentric-milestones.svg` — a work-specific milestone poster generated from the `milestone` and `milestone_image` fields in `data/resources.yml`. The factual labels, dates, and categories are vector text, while the visual panels come from the local `milestones/*.png` images and are rendered uncropped.
 
 When the catalog changes materially, refresh these figures so they stay in sync with `data/resources.yml` (recompute counts, then update the bar geometry, labels, and milestone cards).
 

awesome-egocentric-atlas.csv

@@ -1,6 +1,25 @@
 name,kind,released,venue,status,scope,year,url,paper,code,license,scale,tasks,modalities
 Xperience-10M,dataset,2026-03,Hugging Face,request,,2026,https://huggingface.co/datasets/ropedia-ai/xperience-10m,,,other,"10M experiences, 10K hours, six video streams, audio, stereo depth, camera pose, hand/body mocap, IMU, hierarchical language, about 1 PB total",embodied-ai; world-modeling; robot-learning; sensor-fusion; 3d-4d-understanding; imitation-learning,fisheye-video; stereo-video; audio; stereo-depth; camera-pose; hand-mocap; full-body-mocap; imu; hierarchical-language
 Xperience-10M Sample,dataset,2026-03,Hugging Face,open,,2026,https://huggingface.co/datasets/ropedia-ai/xperience-10m-sample,,,cc-by-nc-4.0,Public sample episode; Hugging Face viewer reports 6 rows,sample-data; loader-testing; visualization; task-suite-prototyping,video; hdf5-annotations
+HALOMI,model,2026-06,arXiv,watch,,2026,https://arxiv.org/abs/2606.18772,https://arxiv.org/abs/2606.18772,,,"Humanoid loco-manipulation from human demonstrations, extending UMI-style collection with egocentric head/wrist observations and head-hand trajectories",humanoid-control; manipulation; imitation-learning; active-perception; vla,egocentric-video; wrist-camera-video; human-demonstrations; robot-actions; humanoid-motion
+HumanoidArena,benchmark,2026-06,arXiv,watch,,2026,https://arxiv.org/abs/2606.17833,https://arxiv.org/abs/2606.17833,,,Benchmark for egocentric hierarchical whole-body learning with seven leg-critical humanoid-object and humanoid-scene interaction tasks,benchmark; humanoid-control; embodied-ai; manipulation; vla,egocentric-video; text-instructions; robot-actions; humanoid-motion
+VLESA,benchmark,2026-06,arXiv,watch,,2026,https://github.com/HanjiangHu/VLESA,https://arxiv.org/abs/2606.03954,https://github.com/HanjiangHu/VLESA,,Egocentric-video safety-assistance benchmark with goal-conditioned safety annotations for human activity monitoring,safety; activity-recognition; video-language; situated-assistance,egocentric-video; annotations; text
+OVO-S-Bench,benchmark,2026-06,arXiv,watch,,2026,https://arxiv.org/abs/2606.03890,https://arxiv.org/abs/2606.03890,,,"1,680 spatial-reasoning questions over 348 continuous egocentric videos with query timestamps and evidence intervals",spatial-reasoning; egocentric-video-qa; streaming-video-understanding; spatial-localization,egocentric-video; text; temporal-grounding; annotations
+EgoPro-Bench,benchmark,2026-05,arXiv,watch,,2026,https://arxiv.org/abs/2605.07299,https://arxiv.org/abs/2605.07299,,,"Personalized proactive-interaction benchmark with 2,400 evaluation videos, 12K+ training videos, and 12 domains",proactive-assistance; personalized-qa; streaming-video-understanding; intent-understanding,egocentric-video; text; annotations
+Pro2Assist,model,2026-05,arXiv,watch,,2026,https://arxiv.org/abs/2605.04227,https://arxiv.org/abs/2605.04227,,,"Continuous step-aware proactive assistance with multimodal egocentric perception, including public-source data and an AR-glasses testbed",proactive-assistance; step-grounding; real-time-assistance; procedure-understanding,egocentric-video; smart-glasses; text; action-labels
+EgoSPT / SPOT,dataset,2026-05,arXiv,watch,,2026,https://arxiv.org/abs/2605.20085,https://arxiv.org/abs/2605.20085,,,Egocentric spatially prompted manipulation trajectories with first-frame object/target grounding and 3D end-effector motion,trajectory-prediction; visual-grounding; manipulation; robot-learning; vla,egocentric-video; trajectories; object-masks; 3d-pose; text-instructions
+EgoBabyVLM,benchmark,2026-05,arXiv,watch,,2026,https://arxiv.org/abs/2605.19130,https://arxiv.org/abs/2605.19130,,,"Benchmarking cross-modal learning from naturalistic infant and adult egocentric video, including Machine-DevBench and an EgoBabyVLM challenge",child-view-learning; video-language; representation-learning; vlm-evaluation,child-view-video; egocentric-video; audio; language
+EgoInteract,dataset,2026-05,arXiv,watch,,2026,https://arxiv.org/abs/2605.18214,https://arxiv.org/abs/2605.18214,,,Controllable egocentric-video simulator and synthetic dataset with dense spatial and temporal annotations,egocentric-simulation; temporal-segmentation; anticipation; next-active-object; ehoi,synthetic-egocentric; egocentric-video; annotations; scene-graphs
+EggHand,model,2026-05,CVPR 2026 Findings,watch,,2026,https://jyoun9.github.io/EggHand/,https://arxiv.org/abs/2605.07642,,,Multimodal foundation model for egocentric hand-pose forecasting over EgoExo4D-style video-language and action signals,hand-forecasting; 3d-hand-pose; video-language; vla,egocentric-video; hand-pose; language; action-labels
+Map-Mono-Ego,model,2026-05,arXiv,watch,,2026,https://arxiv.org/abs/2605.20889,https://arxiv.org/abs/2605.20889,,,"Map-grounded global human-pose estimation from monocular egocentric video, with AIST-Living paired ego video and scanned environments",3d-human-pose; scene-understanding; localization; pose-and-body,egocentric-video; human-pose; scene-scans; camera-trajectory
+Being-H0.7,model,2026-05,arXiv,watch,,2026,https://arxiv.org/abs/2605.00078,https://arxiv.org/abs/2605.00078,,,Latent world-action model learned from egocentric videos for future-aware reasoning and VLA policy learning,world-modeling; vla; robot-learning; future-state-prediction,egocentric-video; latent-actions; robot-actions
+GazeVLA,model,2026-04,arXiv,watch,,2026,https://gazevla.github.io/,https://arxiv.org/abs/2604.22615,,,"Vision-language-action policy that pretrains on large-scale egocentric human data to capture gaze, intention, and action before robot fine-tuning",vla; gaze; intent-understanding; manipulation; robot-learning,egocentric-video; gaze; language; robot-actions
+WARPED,model,2026-04,arXiv,watch,,2026,https://arxiv.org/abs/2604.10809,https://arxiv.org/abs/2604.10809,,,Wrist-aligned rendering pipeline that converts monocular egocentric human demonstrations into robot policy observations with 3D Gaussian Splatting,imitation-learning; cross-embodiment-transfer; robot-learning; manipulation,egocentric-video; wrist-camera-video; 3d-gaussian-splatting; hand-trajectories
+PIE-V,benchmark,2026-04,arXiv,watch,,2026,https://arxiv.org/abs/2604.15134,https://arxiv.org/abs/2604.15134,,,Mistake-aware procedural egocentric-video benchmark injecting human-plausible mistakes and recovery corrections across Ego-Exo4D scenarios,mistake-detection; procedural-reasoning; action-and-procedure; procedural-assistance,egocentric-video; annotations; structured-reasoning
+DP-DeGauss,model,2026-04,arXiv,watch,,2026,https://arxiv.org/abs/2604.07986,https://arxiv.org/abs/2604.07986,,,"Dynamic probabilistic Gaussian decomposition for egocentric 4D scene reconstruction, separating background, hands, and objects in first-person interaction",4d-reconstruction; scene-reconstruction; hand-object-interaction; neural-scene-reconstruction,egocentric-video; 3d-gaussian-splatting; hand-object-motion; 4d-scene
+LifeEval,benchmark,2026-03,arXiv,watch,,2026,https://arxiv.org/abs/2603.00490,https://arxiv.org/abs/2603.00490,,,"4,075 high-quality QA pairs over continuous first-person streams for real-time task-oriented human-AI collaboration in daily life",egocentric-video-qa; interactive-assistance; real-time-assistance; multimodal-reasoning,egocentric-video; dialogue; text
+SAVA-X,benchmark,2026-03,CVPR 2026,watch,,2026,https://arxiv.org/abs/2603.12764,https://arxiv.org/abs/2603.12764,,,"Ego-to-exo imitation-error detection over asynchronous, length-mismatched egocentric and exocentric videos, evaluated with EgoMe",mistake-detection; cross-view; imitation-learning; procedural-assistance,egocentric-video; exocentric-video; annotations
+AG-EgoPose,model,2026-03,arXiv,watch,,2026,https://arxiv.org/abs/2603.25175,https://arxiv.org/abs/2603.25175,,,Attention-guided egocentric 3D human-pose estimation from fisheye camera input with dual motion/spatial streams,egocentric-3d-pose; 3d-human-pose; pose-and-body; motion,fisheye-video; 3d-human-pose; motion
 Ego4D,dataset,2021-10,CVPR 2022,request,,2021,https://ego4d-data.org/,https://arxiv.org/abs/2110.07058,https://github.com/facebookresearch/Ego4d,Ego4D License Agreement,"3,670+ hours, 900+ camera wearers, 74 locations, 9 countries",episodic-memory; nlq; moment-query; hand-object; forecasting; social; audio-visual,video; audio; gaze; stereo; 3d; narrations
 Ego-Exo4D,dataset,2023-11,CVPR 2024,request,,2023,https://ego-exo4d-data.org/,https://arxiv.org/abs/2311.18259,https://github.com/facebookresearch/projectaria_tools,Ego-Exo4D License Agreement,"1,286 hours, 740 participants, synchronized ego/exo views",ego-exo; skilled-activity; cross-view; proficiency; pose,video; audio; gaze; imu; 3d-point-clouds; camera-poses; language
 EPIC-KITCHENS-100,dataset,2020-06,IJCV 2022,open,,2020,https://epic-kitchens.github.io/,https://arxiv.org/abs/2006.13256,https://github.com/epic-kitchens/epic-kitchens-100-annotations,CC BY-NC 4.0,"100 hours, 20M frames, 90K action segments, 45 kitchens",action-recognition; action-detection; anticipation; retrieval; domain-adaptation,video; audio; narrations; object-boxes
@@ -58,7 +77,7 @@ EgoClip,dataset,2022-06,NeurIPS 2022,open,,2022,https://github.com/showlab/EgoVL
 RefEgo,benchmark,2023-08,ICCV 2023,open,,2023,https://github.com/shuheikurita/RefEgo,https://arxiv.org/abs/2308.12035,,,12K+ clips and 41 hours for video-based referring-expression comprehension,referring-expression-comprehension; object-grounding; referred-object-tracking,
 EgoBench,benchmark,2026-05,arXiv,watch,,2026,https://arxiv.org/abs/2605.27820,https://arxiv.org/abs/2605.27820,,,"1,045 egocentric-video-grounded interactive tasks",tool-using-agents; multimodal-reasoning; interaction,
 EgoIntrospect,dataset,2026-05,arXiv,watch,,2026,https://ego-introspect.github.io/,https://arxiv.org/abs/2605.17262,,,"180 hours, 60 subjects",internal-state-reasoning; affect; intent; memory,video; audio; gaze; motion; physiological-signals; self-annotations
-Minerva-Ego,benchmark,2025-05,arXiv,open,adjacent,2025,https://github.com/google-deepmind/neptune,https://arxiv.org/abs/2505.00681,,,Part of the Neptune / MINERVA long-video reasoning collection over web (YouTube) videos with manually annotated reasoning traces,multistep-reasoning; long-video-qa; reasoning-traces,
+Minerva-Ego,benchmark,2026-05,arXiv,open,,2026,https://github.com/google-deepmind/neptune,https://arxiv.org/abs/2605.15342,,,"Complex egocentric visual-reasoning benchmark with multi-step multimodal questions, dense reasoning traces, and spatiotemporal object masks",multistep-reasoning; egocentric-video-qa; spatial-reasoning; visual-grounding; reasoning-traces,egocentric-video; text; segmentation; structured-reasoning
 MA-EgoQA,benchmark,2026-03,arXiv,open,,2026,https://ma-egoqa.github.io/,https://arxiv.org/abs/2603.09827,,,1.7K questions over multiple egocentric streams,multi-agent-memory; social-reasoning; temporal-reasoning,
 EASG-Bench,benchmark,2025-06,ICCV 2025 Workshop,open,,2025,https://github.com/fpv-iplab/EASG-bench,https://arxiv.org/abs/2506.05787,,,,scene-graph-qa; relation-reasoning; temporal-reasoning,
 MyEgo,benchmark,2026-04,CVPR 2026,open,,2026,https://github.com/Ryougetsu3606/MyEgo,https://arxiv.org/abs/2604.01966,,,"541 long videos, 5K personalized questions",personalized-qa; ego-grounding; long-range-memory,

data/resources.yml

@@ -1,7 +1,7 @@
 meta:
   title: Awesome Egocentric Atlas
   description: "Just launched Awesome Egocentric Atlas — a curated collection of egocentric AI resources for egocentric vision, embodied AI, robotics, VLA, world models, WMA, memory, AR/VR, and hand-object interaction."
-  last_major_audit: "2026-06-20"
+  last_major_audit: "2026-06-21"
   scope: Egocentric, first-person, wearable-camera, AR/VR headset, body/wrist camera, and ego-exo datasets, benchmarks, models, and tools.
   status_legend:
     open: Public download, public annotations, public code, or application-based access is clearly documented.
@@ -16,6 +16,7 @@ resources:
     milestone: "2026-03"
     milestone_note: "Petascale egocentric world-model corpus (10M experiences, ~1 PB) pushing egocentric data to internet scale for embodied AI and robot learning."
     milestone_image: assets/milestones/xperience-10m.png
+    milestone_origin: Ropedia
     kind: dataset
     released: "2026-03"
     venue: "Hugging Face"
@@ -46,10 +47,259 @@ resources:
     license: "cc-by-nc-4.0"
     license_url: "https://huggingface.co/datasets/ropedia-ai/xperience-10m-sample"
     verified_at: "2026-06-20"
+  - name: HALOMI
+    kind: model
+    released: "2026-06"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2606.18772
+    paper: https://arxiv.org/abs/2606.18772
+    scale: "Humanoid loco-manipulation from human demonstrations, extending UMI-style collection with egocentric head/wrist observations and head-hand trajectories"
+    modalities: [egocentric-video, wrist-camera-video, human-demonstrations, robot-actions, humanoid-motion]
+    tasks: [humanoid-control, manipulation, imitation-learning, active-perception, vla]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: HumanoidArena
+    kind: benchmark
+    released: "2026-06"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2606.17833
+    paper: https://arxiv.org/abs/2606.17833
+    scale: "Benchmark for egocentric hierarchical whole-body learning with seven leg-critical humanoid-object and humanoid-scene interaction tasks"
+    modalities: [egocentric-video, text-instructions, robot-actions, humanoid-motion]
+    tasks: [benchmark, humanoid-control, embodied-ai, manipulation, vla]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: VLESA
+    kind: benchmark
+    released: "2026-06"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://github.com/HanjiangHu/VLESA
+    paper: https://arxiv.org/abs/2606.03954
+    scale: "Egocentric-video safety-assistance benchmark with goal-conditioned safety annotations for human activity monitoring"
+    modalities: [egocentric-video, annotations, text]
+    tasks: [safety, activity-recognition, video-language, situated-assistance]
+    code: https://github.com/HanjiangHu/VLESA
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: OVO-S-Bench
+    kind: benchmark
+    released: "2026-06"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2606.03890
+    paper: https://arxiv.org/abs/2606.03890
+    scale: "1,680 spatial-reasoning questions over 348 continuous egocentric videos with query timestamps and evidence intervals"
+    modalities: [egocentric-video, text, temporal-grounding, annotations]
+    tasks: [spatial-reasoning, egocentric-video-qa, streaming-video-understanding, spatial-localization]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: EgoPro-Bench
+    kind: benchmark
+    released: "2026-05"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2605.07299
+    paper: https://arxiv.org/abs/2605.07299
+    scale: "Personalized proactive-interaction benchmark with 2,400 evaluation videos, 12K+ training videos, and 12 domains"
+    modalities: [egocentric-video, text, annotations]
+    tasks: [proactive-assistance, personalized-qa, streaming-video-understanding, intent-understanding]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: Pro2Assist
+    kind: model
+    released: "2026-05"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2605.04227
+    paper: https://arxiv.org/abs/2605.04227
+    scale: "Continuous step-aware proactive assistance with multimodal egocentric perception, including public-source data and an AR-glasses testbed"
+    modalities: [egocentric-video, smart-glasses, text, action-labels]
+    tasks: [proactive-assistance, step-grounding, real-time-assistance, procedure-understanding]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: EgoSPT / SPOT
+    kind: dataset
+    released: "2026-05"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2605.20085
+    paper: https://arxiv.org/abs/2605.20085
+    scale: "Egocentric spatially prompted manipulation trajectories with first-frame object/target grounding and 3D end-effector motion"
+    modalities: [egocentric-video, trajectories, object-masks, 3d-pose, text-instructions]
+    tasks: [trajectory-prediction, visual-grounding, manipulation, robot-learning, vla]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: EgoBabyVLM
+    kind: benchmark
+    released: "2026-05"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2605.19130
+    paper: https://arxiv.org/abs/2605.19130
+    scale: "Benchmarking cross-modal learning from naturalistic infant and adult egocentric video, including Machine-DevBench and an EgoBabyVLM challenge"
+    modalities: [child-view-video, egocentric-video, audio, language]
+    tasks: [child-view-learning, video-language, representation-learning, vlm-evaluation]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: EgoInteract
+    kind: dataset
+    released: "2026-05"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2605.18214
+    paper: https://arxiv.org/abs/2605.18214
+    scale: "Controllable egocentric-video simulator and synthetic dataset with dense spatial and temporal annotations"
+    modalities: [synthetic-egocentric, egocentric-video, annotations, scene-graphs]
+    tasks: [egocentric-simulation, temporal-segmentation, anticipation, next-active-object, ehoi]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: EggHand
+    kind: model
+    released: "2026-05"
+    venue: "CVPR 2026 Findings"
+    year: 2026
+    status: watch
+    url: https://jyoun9.github.io/EggHand/
+    paper: https://arxiv.org/abs/2605.07642
+    scale: "Multimodal foundation model for egocentric hand-pose forecasting over EgoExo4D-style video-language and action signals"
+    modalities: [egocentric-video, hand-pose, language, action-labels]
+    tasks: [hand-forecasting, 3d-hand-pose, video-language, vla]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: Map-Mono-Ego
+    kind: model
+    released: "2026-05"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2605.20889
+    paper: https://arxiv.org/abs/2605.20889
+    scale: "Map-grounded global human-pose estimation from monocular egocentric video, with AIST-Living paired ego video and scanned environments"
+    modalities: [egocentric-video, human-pose, scene-scans, camera-trajectory]
+    tasks: [3d-human-pose, scene-understanding, localization, pose-and-body]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: Being-H0.7
+    kind: model
+    released: "2026-05"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2605.00078
+    paper: https://arxiv.org/abs/2605.00078
+    scale: "Latent world-action model learned from egocentric videos for future-aware reasoning and VLA policy learning"
+    modalities: [egocentric-video, latent-actions, robot-actions]
+    tasks: [world-modeling, vla, robot-learning, future-state-prediction]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: GazeVLA
+    kind: model
+    released: "2026-04"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://gazevla.github.io/
+    paper: https://arxiv.org/abs/2604.22615
+    scale: "Vision-language-action policy that pretrains on large-scale egocentric human data to capture gaze, intention, and action before robot fine-tuning"
+    modalities: [egocentric-video, gaze, language, robot-actions]
+    tasks: [vla, gaze, intent-understanding, manipulation, robot-learning]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: WARPED
+    kind: model
+    released: "2026-04"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2604.10809
+    paper: https://arxiv.org/abs/2604.10809
+    scale: "Wrist-aligned rendering pipeline that converts monocular egocentric human demonstrations into robot policy observations with 3D Gaussian Splatting"
+    modalities: [egocentric-video, wrist-camera-video, 3d-gaussian-splatting, hand-trajectories]
+    tasks: [imitation-learning, cross-embodiment-transfer, robot-learning, manipulation]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: PIE-V
+    kind: benchmark
+    released: "2026-04"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2604.15134
+    paper: https://arxiv.org/abs/2604.15134
+    scale: "Mistake-aware procedural egocentric-video benchmark injecting human-plausible mistakes and recovery corrections across Ego-Exo4D scenarios"
+    modalities: [egocentric-video, annotations, structured-reasoning]
+    tasks: [mistake-detection, procedural-reasoning, action-and-procedure, procedural-assistance]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: DP-DeGauss
+    kind: model
+    released: "2026-04"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2604.07986
+    paper: https://arxiv.org/abs/2604.07986
+    scale: "Dynamic probabilistic Gaussian decomposition for egocentric 4D scene reconstruction, separating background, hands, and objects in first-person interaction"
+    modalities: [egocentric-video, 3d-gaussian-splatting, hand-object-motion, 4d-scene]
+    tasks: [4d-reconstruction, scene-reconstruction, hand-object-interaction, neural-scene-reconstruction]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: LifeEval
+    kind: benchmark
+    released: "2026-03"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2603.00490
+    paper: https://arxiv.org/abs/2603.00490
+    scale: "4,075 high-quality QA pairs over continuous first-person streams for real-time task-oriented human-AI collaboration in daily life"
+    modalities: [egocentric-video, dialogue, text]
+    tasks: [egocentric-video-qa, interactive-assistance, real-time-assistance, multimodal-reasoning]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: SAVA-X
+    kind: benchmark
+    released: "2026-03"
+    venue: "CVPR 2026"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2603.12764
+    paper: https://arxiv.org/abs/2603.12764
+    scale: "Ego-to-exo imitation-error detection over asynchronous, length-mismatched egocentric and exocentric videos, evaluated with EgoMe"
+    modalities: [egocentric-video, exocentric-video, annotations]
+    tasks: [mistake-detection, cross-view, imitation-learning, procedural-assistance]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
+  - name: AG-EgoPose
+    kind: model
+    released: "2026-03"
+    venue: "arXiv"
+    year: 2026
+    status: watch
+    url: https://arxiv.org/abs/2603.25175
+    paper: https://arxiv.org/abs/2603.25175
+    scale: "Attention-guided egocentric 3D human-pose estimation from fisheye camera input with dual motion/spatial streams"
+    modalities: [fisheye-video, 3d-human-pose, motion]
+    tasks: [egocentric-3d-pose, 3d-human-pose, pose-and-body, motion]
+    verified_at: "2026-06-21"
+    release_note: "Live source checked 2026-06-21; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
   - name: Ego4D
     milestone: "2021-10"
     milestone_note: "The 3,670-hour massive-scale benchmark suite that catalyzed the modern egocentric era."
     milestone_image: assets/milestones/ego4d.png
+    milestone_origin: Meta AI / Ego4D consortium
     kind: dataset
     released: "2021-10"
     venue: "CVPR 2022"
@@ -68,6 +318,7 @@ resources:
     milestone: "2023-11"
     milestone_note: "Synchronized ego and exo skilled-activity capture at scale; the reference for cross-view egocentric learning."
     milestone_image: assets/milestones/ego-exo4d.png
+    milestone_origin: Meta AI / Ego-Exo4D consortium
     kind: dataset
     released: "2023-11"
     venue: "CVPR 2024"
@@ -86,6 +337,7 @@ resources:
     milestone: "2020-06"
     milestone_note: "The defining large-scale egocentric action-recognition benchmark and annual challenge suite."
     milestone_image: assets/milestones/epic-kitchens-100.png
+    milestone_origin: Univ. Bristol / Univ. Catania
     kind: dataset
     released: "2020-06"
     venue: "IJCV 2022"
@@ -258,6 +510,7 @@ resources:
     milestone: "2024-02"
     milestone_note: "Handheld/wrist-view manipulation interface that made robot-free in-the-wild demonstrations practical for cross-embodiment policy learning."
     milestone_image: assets/milestones/umi.png
+    milestone_origin: Stanford
     citation_key: umi_2024
 
   - name: FastUMI
@@ -427,6 +680,7 @@ resources:
     milestone: "2024-06"
     milestone_note: "Reference benchmark for 3D hand-object tracking from AR glasses (Project Aria and Quest 3)."
     milestone_image: assets/milestones/hot3d.png
+    milestone_origin: Meta / Project Aria
     kind: dataset
     released: "2024-06"
     venue: "CVPR 2025"
@@ -458,6 +712,7 @@ resources:
     milestone: "2022-03"
     milestone_note: "Large RGB-D 4D hand-object interaction dataset that moved egocentric HOI toward geometry, pose, and temporal scene understanding."
     milestone_image: assets/milestones/hoi4d.png
+    milestone_origin: Tsinghua / Peking / Qi Zhi
 
   - name: H2O
     kind: dataset
@@ -500,6 +755,7 @@ resources:
     milestone: "2015-12"
     milestone_note: "The Google Glass hand-segmentation benchmark that made hands a first-class signal for egocentric perception."
     milestone_image: assets/milestones/egohands.png
+    milestone_origin: Indiana University
 
   - name: FPHA
     kind: dataset
@@ -659,6 +915,7 @@ resources:
     milestone: "2023-08"
     milestone_note: "Meta's research smart-glasses platform that opened the modern wave of AR and wearable egocentric data."
     milestone_image: assets/milestones/project-aria.png
+    milestone_origin: Meta Reality Labs
     kind: collection
     released: "2023-08"
     venue: "arXiv"
@@ -741,6 +998,7 @@ resources:
     milestone: "2025-03"
     milestone_note: "Week-long Meta Aria daily-life corpus that pushed egocentric research toward personalized memory and life-assistant reasoning."
     milestone_image: assets/milestones/egolife.png
+    milestone_origin: S-Lab, NTU / LMMs-Lab
 
   - name: Ego-EXTRA
     kind: dataset
@@ -760,6 +1018,7 @@ resources:
     milestone: "2023-08"
     milestone_note: "The benchmark that exposed how far models are from long-form egocentric video reasoning."
     milestone_image: assets/milestones/egoschema.png
+    milestone_origin: UC Berkeley
     kind: benchmark
     released: "2023-08"
     venue: "NeurIPS 2023"
@@ -868,15 +1127,17 @@ resources:
     release_note: "Live source checked 2026-06-20; kept on watch until public artifacts, access terms, or reproducible release metadata are confirmed."
   - name: Minerva-Ego
     kind: benchmark
-    released: "2025-05"
+    released: "2026-05"
     venue: "arXiv"
-    year: 2025
+    year: 2026
     status: open
-    scope: adjacent
     url: https://github.com/google-deepmind/neptune
-    paper: https://arxiv.org/abs/2505.00681
-    scale: "Part of the Neptune / MINERVA long-video reasoning collection over web (YouTube) videos with manually annotated reasoning traces"
-    tasks: [multistep-reasoning, long-video-qa, reasoning-traces]
+    paper: https://arxiv.org/abs/2605.15342
+    scale: "Complex egocentric visual-reasoning benchmark with multi-step multimodal questions, dense reasoning traces, and spatiotemporal object masks"
+    modalities: [egocentric-video, text, segmentation, structured-reasoning]
+    tasks: [multistep-reasoning, egocentric-video-qa, spatial-reasoning, visual-grounding, reasoning-traces]
+    verified_at: "2026-06-21"
+    release_note: "Paper points to the google-deepmind/neptune repository for download; release state should be rechecked as the benchmark matures."
 
   - name: MA-EgoQA
     kind: benchmark
@@ -1097,6 +1358,7 @@ resources:
     milestone: "2011-06"
     milestone_note: "Foundational hand-object and gaze egocentric activity datasets (GTEA, CVPR 2011) that seeded egocentric action and attention research."
     milestone_image: assets/milestones/gtea-gaze.png
+    milestone_origin: Georgia Tech
     kind: dataset
     released: "2018"
     venue: "project page"
@@ -1264,6 +1526,7 @@ resources:
     milestone: "2022-06"
     milestone_note: "First egocentric video-language pretraining (EgoClip, EgoNCE) and a basis for ego representation learning."
     milestone_image: assets/milestones/egovlp.png
+    milestone_origin: ShowLab
     kind: model
     released: "2022-06"
     venue: "NeurIPS 2022"
@@ -2136,6 +2399,7 @@ resources:
     milestone: "2025-07"
     milestone_note: "Showed vision-language-action policies can be learned from egocentric human video and transferred to robots."
     milestone_image: assets/milestones/egovla.png
+    milestone_origin: UCSD / UIUC / MIT / NVIDIA
     kind: model
     released: "2025-07"
     venue: "project page"
@@ -2305,6 +2569,7 @@ resources:
     milestone: "2012-06"
     milestone_note: "Classic unscripted daily-life egocentric activity corpus with object and hand annotations; the reference point for ADL-scale egocentric recognition."
     milestone_image: assets/milestones/adl-dataset.png
+    milestone_origin: UC Irvine
 
   - name: Wrist-mounted ADL
     kind: dataset
@@ -2996,6 +3261,7 @@ resources:
     milestone: "2026-02"
     milestone_note: "Revealed the log-linear data-scaling law for egocentric human-video VLA pretraining."
     milestone_image: assets/milestones/egoscale.png
+    milestone_origin: NVIDIA / UC Berkeley / UMD
     kind: dataset
     released: "2026-02"
     venue: "arXiv"
@@ -4322,6 +4588,7 @@ resources:
     milestone: "2009-06"
     milestone_note: "The earliest egocentric dataset; launched egocentric activity recognition at the first IEEE Workshop on Egocentric Vision (CVPR 2009)."
     milestone_image: assets/milestones/cmu-mmac.png
+    milestone_origin: Carnegie Mellon University
     kind: dataset
     released: "2009-06"
     venue: "CMU tech report 2009"
@@ -4384,6 +4651,7 @@ resources:
     milestone: "2026-02"
     milestone_note: "44K-hour egocentric-video robot world model with latent actions for real-time planning."
     milestone_image: assets/milestones/dreamdojo.png
+    milestone_origin: NVIDIA
     citation_key: dreamdojo_2026
 
   - name: UniDex

docs/resource_schema.md

@@ -55,6 +55,7 @@ enforces the required fields and the format of the optional ones it understands.
 | `released_by` | string | Releasing org/channel. |
 | `publisher` | string | Hosting platform. |
 | `derived_from` | string | Parent dataset this is built on. |
+| `milestone_origin` | string | Short institution, lab, company, project, or consortium label shown on milestone cards when the entry has `milestone`. |
 
 ## Minimal example
 

index.html

@@ -65,6 +65,20 @@
       "keywords": "sample-data, loader-testing, visualization, task-suite-prototyping",
       "license": "https://huggingface.co/datasets/ropedia-ai/xperience-10m-sample"
     },
+    {
+      "@type": "Dataset",
+      "name": "EgoSPT / SPOT",
+      "url": "https://arxiv.org/abs/2605.20085",
+      "description": "Egocentric spatially prompted manipulation trajectories with first-frame object/target grounding and 3D end-effector motion",
+      "keywords": "trajectory-prediction, visual-grounding, manipulation, robot-learning, vla"
+    },
+    {
+      "@type": "Dataset",
+      "name": "EgoInteract",
+      "url": "https://arxiv.org/abs/2605.18214",
+      "description": "Controllable egocentric-video simulator and synthetic dataset with dense spatial and temporal annotations",
+      "keywords": "egocentric-simulation, temporal-segmentation, anticipation, next-active-object, ehoi"
+    },
     {
       "@type": "Dataset",
       "name": "Ego4D",
@@ -940,14 +954,6 @@
     <link rel="icon" href="assets/awesome-egocentric-logo.png" type="image/png">
     <link rel="apple-touch-icon" href="assets/awesome-egocentric-logo.png">
     <link rel="stylesheet" href="./styles.css">
-    <style>
-      .site-footer { display: flex; flex-direction: column; gap: 0.85rem; text-align: center; align-items: center; }
-      .lang-bar { font-size: 0.85rem; line-height: 2; color: #64748b; }
-      .lang-bar a { color: #2563eb; text-decoration: none; }
-      .lang-bar a:hover { text-decoration: underline; }
-      .lang-bar a.active { font-weight: 700; color: #0f3b45; text-decoration: underline; }
-      .footer-meta { display: flex; flex-wrap: wrap; gap: 0.4rem 1.2rem; justify-content: center; align-items: center; }
-    </style>
   </head>
   <body>
     <a class="skip-link" href="#catalog" data-i18n="skip">Skip to catalog</a>
@@ -963,6 +969,15 @@
         <a href="#access" data-i18n="nav.access">Access</a>
         <a href="https://github.com/ChaoYue0307/awesome-egocentric-atlas#readme" data-i18n="nav.readme">README</a>
       </nav>
+      <div class="header-actions">
+        <details class="language-switcher" id="language-switcher">
+          <summary class="language-toggle" aria-label="Select language">
+            <span class="language-label" data-i18n="language.label">Language</span>
+            <span class="language-current" id="language-current">English</span>
+          </summary>
+          <nav class="language-menu" id="language-menu" aria-label="Choose language"></nav>
+        </details>
+      </div>
     </header>
 
     <main>
@@ -977,19 +992,19 @@
           </div>
           <dl class="stat-strip" aria-label="Atlas summary">
             <div>
-              <dt data-stat="egocentric_resources">456</dt>
+              <dt data-stat="egocentric_resources">476</dt>
               <dd data-i18n="stat.resources">egocentric resources</dd>
             </div>
             <div>
-              <dt data-kind="dataset">125</dt>
+              <dt data-kind="dataset">127</dt>
               <dd data-i18n="stat.datasets">datasets</dd>
             </div>
             <div>
-              <dt data-kind="benchmark">81</dt>
+              <dt data-kind="benchmark">90</dt>
               <dd data-i18n="stat.benchmarks">benchmarks</dd>
             </div>
             <div>
-              <dt data-kind="model">226</dt>
+              <dt data-kind="model">235</dt>
               <dd data-i18n="stat.models">models</dd>
             </div>
             <div>
@@ -1007,7 +1022,7 @@
           <img src="assets/awesome-egocentric-atlas-cover.png" width="1672" height="941" decoding="async" fetchpriority="high" alt="Technical illustration of a first-person egocentric AI atlas">
           <div class="media-caption">
             <span data-i18n="media.caption">Egocentric AI, mapped</span>
-            <span data-updated>Updated 2026-06-20</span>
+            <span data-updated>Updated 2026-06-21</span>
           </div>
         </div>
       </section>
@@ -1023,12 +1038,7 @@
             <p data-i18n="milestones.desc">The landmark works that shaped egocentric AI — from the field's origins to its current frontier.</p>
           </div>
         </div>
-        <figure class="milestone-poster">
-          <div class="milestone-poster-frame">
-            <img src="assets/awesome-egocentric-milestones.png" width="3840" height="6966" loading="lazy" decoding="async" alt="Illustrated milestone timeline for representative egocentric AI works">
-            <div class="milestone-link-layer" id="milestone-link-layer" aria-label="Milestone work links"></div>
-          </div>
-        </figure>
+        <div class="milestone-board" id="milestone-board" aria-label="Representative milestone works"></div>
       </section>
 
       <section class="section" id="catalog">
@@ -1142,7 +1152,6 @@
     </main>
 
     <footer class="site-footer">
-      <nav class="lang-bar" id="lang-bar" aria-label="Language"></nav>
       <div class="footer-meta">
         <span>Awesome Egocentric Atlas</span>
         <span data-i18n="footer.tagline">MIT licensed and free to use. Contributions welcome.</span>

scripts/build_site_data.rb

@@ -1,7 +1,7 @@
 #!/usr/bin/env ruby
 # frozen_string_literal: true
 
-# Backward-compatible wrapper for the site JSON/CSV subset of the artifact build.
+# Backward-compatible wrapper for the site JSON/CSV export subset of the artifact build.
 #
 # Usage:
 #   ruby scripts/build_site_data.rb
@@ -11,7 +11,8 @@ require_relative "lib/catalog_artifacts"
 
 targets = {
   CatalogArtifacts::SITE_DATA => CatalogArtifacts.site_json,
-  CatalogArtifacts::CSV_OUTPUT => CatalogArtifacts.csv
+  CatalogArtifacts::CSV_OUTPUT => CatalogArtifacts.csv,
+  CatalogArtifacts::PAPERS_CSV_OUTPUT => CatalogArtifacts.papers_csv
 }
 
 if ARGV.include?("--check")
@@ -33,5 +34,5 @@ if ARGV.include?("--check")
   puts "Site data OK: #{summary.fetch('egocentric_resources')} egocentric resources (#{summary.fetch('total_resources')} rows)"
 else
   targets.each { |path, expected| File.write(path, expected, encoding: "UTF-8") }
-  puts "Wrote #{CatalogArtifacts::SITE_DATA} and #{CatalogArtifacts::CSV_OUTPUT}"
+  puts "Wrote #{targets.keys.map { |path| path.sub("#{CatalogArtifacts::ROOT}/", "") }.join(', ')}"
 end

scripts/lib/catalog_artifacts.rb

@@ -15,10 +15,12 @@ module CatalogArtifacts
   README_ALIASES = File.join(ROOT, "data", "readme_aliases.yml")
   SITE_DATA = File.join(ROOT, "site-data.json")
   CSV_OUTPUT = File.join(ROOT, "awesome-egocentric-atlas.csv")
+  PAPERS_CSV_OUTPUT = File.join(ROOT, "awesome-egocentric-papers.csv")
   HF_HEADER = File.join(ROOT, "huggingface", "dataset_card_header.txt")
   MILESTONES_SVG = File.join(ROOT, "assets", "awesome-egocentric-milestones.svg")
 
   CSV_COLUMNS = %w[name kind released venue status scope year url paper code license scale tasks modalities].freeze
+  PAPERS_CSV_COLUMNS = %w[name kind released venue status scope year paper url code license scale tasks modalities].freeze
   STATUS_ORDER = %w[open watch partial benchmark request].freeze
   KIND_ORDER = %w[dataset benchmark model toolkit collection].freeze
 
@@ -161,7 +163,8 @@ module CatalogArtifacts
           "kind" => entry["kind"],
           "url" => entry["url"],
           "date" => entry["milestone"].to_s,
-          "note" => entry["milestone_note"]
+          "note" => entry["milestone_note"],
+          "origin" => entry["milestone_origin"] || entry["released_by"] || entry["publisher"] || entry["created_by"] || entry["venue"]
         }
         milestone["image"] = entry["milestone_image"] if entry["milestone_image"]
         milestone
@@ -200,18 +203,29 @@ module CatalogArtifacts
     JSON.pretty_generate(site_payload) + "\n"
   end
 
-  def csv
+  def csv_value(value)
+    value.is_a?(Array) ? value.join("; ") : value
+  end
+
+  def csv_for(rows, columns)
     CSV.generate do |out|
-      out << CSV_COLUMNS
-      resources.each do |entry|
-        out << CSV_COLUMNS.map do |column|
-          value = entry[column]
-          value.is_a?(Array) ? value.join("; ") : value
+      out << columns
+      rows.each do |entry|
+        out << columns.map do |column|
+          csv_value(entry[column])
         end
       end
     end
   end
 
+  def csv
+    csv_for(resources, CSV_COLUMNS)
+  end
+
+  def papers_csv
+    csv_for(site_resources.select { |entry| entry["paper"].to_s.strip != "" }, PAPERS_CSV_COLUMNS)
+  end
+
   def number_word(value)
     words = {
       0 => "zero", 1 => "one", 2 => "two", 3 => "three", 4 => "four",
@@ -788,34 +802,45 @@ module CatalogArtifacts
       }
     ].reject { |group| group.fetch(:items).empty? }
 
-    margin_x = 40
+    canvas_width = 1600
+    margin_x = 52
     row_gap = 26
     header_height = 188
-    row_width = 1200
-    card_area_x = 266
-    card_area_width = 954
-    card_gap = 14
+    row_width = canvas_width - (margin_x * 2)
+    card_area_x = 330
+    card_area_width = canvas_width - card_area_x - margin_x
+    card_gap = 18
+    card_row_gap = 20
     era_dot_x = card_area_x - margin_x - 18
 
     row_layouts = era_specs.map do |group|
       count = group.fetch(:items).length
+      columns = count >= 5 ? 3 : [count, 4].min
       card_width = [
         [
-          ((card_area_width - (card_gap * [count - 1, 0].max)) / count.to_f).floor,
-          174
+          ((card_area_width - (card_gap * [columns - 1, 0].max)) / columns.to_f).floor,
+          252
         ].max,
-        286
+        300
       ].min
       image_height = [
         [
-          (card_width * 0.62).round,
-          104
+          (card_width * 0.60).round,
+          156
         ].max,
-        158
+        180
       ].min
-      card_height = image_height + 154
-      row_height = card_height + 104
-      group.merge(card_width: card_width, image_height: image_height, card_height: card_height, row_height: row_height)
+      card_height = image_height + 262
+      card_rows = (count.to_f / columns).ceil
+      card_stack_height = (card_rows * card_height) + (card_row_gap * [card_rows - 1, 0].max)
+      row_height = card_stack_height + 124
+      group.merge(
+        columns: columns,
+        card_width: card_width,
+        image_height: image_height,
+        card_height: card_height,
+        row_height: row_height
+      )
     end
 
     height = header_height + row_layouts.sum { |group| group.fetch(:row_height) } + (row_gap * [row_layouts.length - 1, 0].max) + 31
@@ -826,20 +851,23 @@ module CatalogArtifacts
       y = current_y
       current_y += group.fetch(:row_height) + row_gap
       era_nodes << [margin_x + era_dot_x, y + 54]
-      count = group.fetch(:items).length
-      cards_width = (count * group.fetch(:card_width)) + (card_gap * [count - 1, 0].max)
-      start_x = card_area_x + ((card_area_width - cards_width) / 2.0)
       card_y = y + 78
 
       cells = group.fetch(:items).each_with_index.map do |item, index|
         card_width = group.fetch(:card_width)
         image_height = group.fetch(:image_height)
         card_height = group.fetch(:card_height)
-        x = start_x + (index * (card_width + card_gap))
+        columns = group.fetch(:columns)
+        row_index = index / columns
+        column_index = index % columns
+        items_in_row = [group.fetch(:items).length - (row_index * columns), columns].min
+        cards_width = (items_in_row * card_width) + (card_gap * [items_in_row - 1, 0].max)
+        start_x = card_area_x + ((card_area_width - cards_width) / 2.0)
+        x = start_x + (column_index * (card_width + card_gap))
         {
           item: item,
           x: x,
-          y: card_y,
+          y: card_y + (row_index * (card_height + card_row_gap)),
           width: card_width,
           height: card_height,
           image_height: image_height
@@ -852,7 +880,7 @@ module CatalogArtifacts
     {
       items: items,
       year_span: year_span,
-      width: 1280,
+      width: canvas_width,
       height: height,
       margin_x: margin_x,
       row_width: row_width,
@@ -890,6 +918,7 @@ module CatalogArtifacts
     layout = milestone_poster_layout
     items = layout.fetch(:items)
     year_span = layout.fetch(:year_span)
+    width = layout.fetch(:width)
     height = layout.fetch(:height)
     margin_x = layout.fetch(:margin_x)
     row_width = layout.fetch(:row_width)
@@ -920,11 +949,16 @@ module CatalogArtifacts
         kind_color = kind_colors.fetch(kind, group.fetch(:accent))
         date_width = [date.length * 7.3 + 22, 66].max.round
         kind_width = [[kind.length * 6.9 + 24, 58].max.round, card_width - date_width - 25].min
-        max_name_chars = [[(card_width / 7.7).floor, 15].max, 30].min
-        max_note_chars = [[(card_width / 5.9).floor, 20].max, 48].min
+        max_name_chars = [[(card_width / 8.5).floor, 18].max, 34].min
+        max_origin_chars = [[(card_width / 6.5).floor, 26].max, 42].min
+        max_note_chars = [[(card_width / 6.9).floor, 25].max, 42].min
         name_lines = wrap_text(item.fetch("name"), max_chars: max_name_chars, max_lines: 2)
-        note_y = image_height + 66 + (name_lines.length * 19) + 9
-        note_lines = wrap_text(item.fetch("note"), max_chars: max_note_chars, max_lines: 2)
+        origin = item["origin"].to_s.strip
+        origin_lines = origin.empty? ? [] : wrap_text("Origin: #{origin}", max_chars: max_origin_chars, max_lines: 1)
+        note_lines = wrap_text(item.fetch("note").to_s, max_chars: max_note_chars, max_lines: 4)
+        title_y = image_height + 84
+        origin_y = title_y + (name_lines.length * 23) + 8
+        note_y = origin_y + (origin_lines.empty? ? 4 : 24)
 
         <<~CARD
           <a href="#{html_escape(item.fetch("url"))}" target="_blank">
@@ -932,12 +966,13 @@ module CatalogArtifacts
               <rect class="card-bg" width="#{card_width}" height="#{card_height}" rx="14"/>
               <rect class="image-frame" x="12" y="12" width="#{card_width - 24}" height="#{image_height}" rx="10"/>
               <image href="#{html_escape(image)}" x="12" y="12" width="#{card_width - 24}" height="#{image_height}" preserveAspectRatio="xMidYMid meet"/>
-              <rect class="date-pill" x="12" y="#{image_height + 23}" width="#{date_width}" height="25" rx="12.5"/>
-              <text class="pill-text" x="#{12 + (date_width / 2.0)}" y="#{image_height + 40}" text-anchor="middle">#{html_escape(date)}</text>
-              <rect class="kind-pill" x="#{19 + date_width}" y="#{image_height + 23}" width="#{kind_width}" height="25" rx="12.5" fill="#{kind_color}"/>
-              <text class="pill-text" x="#{19 + date_width + (kind_width / 2.0)}" y="#{image_height + 40}" text-anchor="middle">#{html_escape(kind)}</text>
-              #{svg_text_block(name_lines, x: 12, y: image_height + 66, class_name: "card-title", line_height: 19)}
-              #{svg_text_block(note_lines, x: 12, y: note_y, class_name: "card-note", line_height: 15)}
+              <rect class="date-pill" x="12" y="#{image_height + 32}" width="#{date_width}" height="27" rx="13.5"/>
+              <text class="pill-text" x="#{12 + (date_width / 2.0)}" y="#{image_height + 50}" text-anchor="middle">#{html_escape(date)}</text>
+              <rect class="kind-pill" x="#{21 + date_width}" y="#{image_height + 32}" width="#{kind_width}" height="27" rx="13.5" fill="#{kind_color}"/>
+              <text class="pill-text" x="#{21 + date_width + (kind_width / 2.0)}" y="#{image_height + 50}" text-anchor="middle">#{html_escape(kind)}</text>
+              #{svg_text_block(name_lines, x: 12, y: title_y, class_name: "card-title", line_height: 23)}
+              #{svg_text_block(origin_lines, x: 12, y: origin_y, class_name: "card-origin", line_height: 17)}
+              #{svg_text_block(note_lines, x: 12, y: note_y, class_name: "card-note", line_height: 17)}
             </g>
           </a>
         CARD
@@ -968,7 +1003,7 @@ module CatalogArtifacts
     end
 
     <<~SVG
-      <svg xmlns="http://www.w3.org/2000/svg" width="1280" height="#{height}" viewBox="0 0 1280 #{height}" role="img" aria-labelledby="title desc">
+      <svg xmlns="http://www.w3.org/2000/svg" width="#{width}" height="#{height}" viewBox="0 0 #{width} #{height}" role="img" aria-labelledby="title desc">
         <title id="title">Representative egocentric AI milestones</title>
         <desc id="desc">A generated milestone poster for Awesome Egocentric Atlas, showing #{items.length} representative field-defining works from #{year_span}, grouped into era bands with uncropped visual panels.</desc>
         <defs>
@@ -990,7 +1025,7 @@ module CatalogArtifacts
           </filter>
           <style>
             .kicker { font: 800 15px system-ui, -apple-system, "Segoe UI", sans-serif; fill: #0b8f98; letter-spacing: .18em; }
-            .title { font: 850 46px system-ui, -apple-system, "Segoe UI", sans-serif; fill: #14212b; letter-spacing: 0; }
+            .title { font: 850 52px system-ui, -apple-system, "Segoe UI", sans-serif; fill: #14212b; letter-spacing: 0; }
             .subtitle { font: 500 18px system-ui, -apple-system, "Segoe UI", sans-serif; fill: #53606b; letter-spacing: 0; }
             .stat-num { font: 900 43px system-ui, -apple-system, "Segoe UI", sans-serif; fill: #0f3b45; letter-spacing: 0; }
             .stat-range { font: 900 34px system-ui, -apple-system, "Segoe UI", sans-serif; fill: #0f3b45; letter-spacing: 0; }
@@ -1010,26 +1045,27 @@ module CatalogArtifacts
             .date-pill { fill: #0b8f98; }
             .kind-pill { fill: #ef9f24; }
             .pill-text { font: 800 10.8px system-ui, -apple-system, "Segoe UI", sans-serif; fill: #ffffff; letter-spacing: 0; }
-            .card-title { font: 820 15px system-ui, -apple-system, "Segoe UI", sans-serif; fill: #182733; letter-spacing: 0; }
-            .card-note { font: 520 11.4px system-ui, -apple-system, "Segoe UI", sans-serif; fill: #5c6b74; letter-spacing: 0; }
+            .card-title { font: 840 18px system-ui, -apple-system, "Segoe UI", sans-serif; fill: #182733; letter-spacing: 0; }
+            .card-origin { font: 780 12.2px system-ui, -apple-system, "Segoe UI", sans-serif; fill: #0b8f98; letter-spacing: .01em; }
+            .card-note { font: 560 12.4px system-ui, -apple-system, "Segoe UI", sans-serif; fill: #5d6c76; letter-spacing: 0; }
             .rule { stroke: #c8dde2; stroke-width: 1.4; }
           </style>
         </defs>
 
-        <rect width="1280" height="#{height}" fill="url(#bg)"/>
-        <rect width="1280" height="#{height}" fill="url(#dots)"/>
-        <rect class="hero-card" x="40" y="24" width="1200" height="138" rx="24"/>
-        <rect x="40" y="24" width="1200" height="138" rx="24" fill="url(#heroGlow)"/>
-        <line class="rule" x1="40" y1="150" x2="1240" y2="150"/>
+        <rect width="#{width}" height="#{height}" fill="url(#bg)"/>
+        <rect width="#{width}" height="#{height}" fill="url(#dots)"/>
+        <rect class="hero-card" x="#{margin_x}" y="24" width="#{row_width}" height="138" rx="24"/>
+        <rect x="#{margin_x}" y="24" width="#{row_width}" height="138" rx="24" fill="url(#heroGlow)"/>
+        <line class="rule" x1="#{margin_x}" y1="150" x2="#{margin_x + row_width}" y2="150"/>
 
-        <text class="kicker" x="40" y="48">CURATED FIELD MILESTONES</text>
-        <text class="title" x="40" y="101">Egocentric AI timeline</text>
-        <text class="subtitle" x="40" y="136">Representative works grouped by the shifts they created: egocentric data, scale, reasoning, robotics, and world models.</text>
+        <text class="kicker" x="#{margin_x}" y="48">CURATED FIELD MILESTONES</text>
+        <text class="title" x="#{margin_x}" y="104">Egocentric AI timeline</text>
+        <text class="subtitle" x="#{margin_x}" y="138">Representative works grouped by the shifts they created: egocentric data, scale, reasoning, robotics, and world models.</text>
 
-        <text class="stat-range" x="1014" y="82" text-anchor="middle">#{year_span}</text>
-        <text class="stat-label" x="1014" y="106" text-anchor="middle">SPAN</text>
-        <text class="stat-num" x="1172" y="82" text-anchor="middle">#{items.length}</text>
-        <text class="stat-label" x="1172" y="106" text-anchor="middle">WORKS</text>
+        <text class="stat-range" x="#{width - 300}" y="82" text-anchor="middle">#{year_span}</text>
+        <text class="stat-label" x="#{width - 300}" y="106" text-anchor="middle">SPAN</text>
+        <text class="stat-num" x="#{width - 104}" y="82" text-anchor="middle">#{items.length}</text>
+        <text class="stat-label" x="#{width - 104}" y="106" text-anchor="middle">WORKS</text>
 
         #{spine_path}
       #{bands}
@@ -1041,6 +1077,7 @@ module CatalogArtifacts
     {
       SITE_DATA => site_json,
       CSV_OUTPUT => csv,
+      PAPERS_CSV_OUTPUT => papers_csv,
       README => updated_readme,
       File.join(ROOT, "index.html") => updated_index,
       File.join(ROOT, "assets", "awesome-egocentric-timeline.svg") => updated_timeline_svg,

scripts/verify_hf_mirror.rb

@@ -73,4 +73,17 @@ unless missing.empty?
   exit 1
 end
 
+local_papers_path = File.join(package_dir, "awesome-egocentric-papers.csv")
+unless File.file?(local_papers_path)
+  warn "HF package is missing awesome-egocentric-papers.csv"
+  exit 1
+end
+
+local_papers = File.read(local_papers_path, encoding: "UTF-8")
+remote_papers = fetch_text("#{base}/awesome-egocentric-papers.csv?download=1&ts=#{Time.now.to_i}")
+unless remote_papers == local_papers
+  warn "HF papers CSV mismatch"
+  exit 1
+end
+
 puts "Hugging Face mirror OK: #{repo_id} (#{local_summary.fetch('egocentric_resources')} egocentric resources, #{local_date})"

site-data.json

@@ -2,7 +2,7 @@
   "meta": {
     "title": "Awesome Egocentric Atlas",
     "description": "Just launched Awesome Egocentric Atlas — a curated collection of egocentric AI resources for egocentric vision, embodied AI, robotics, VLA, world models, WMA, memory, AR/VR, and hand-object interaction.",
-    "last_major_audit": "2026-06-20",
+    "last_major_audit": "2026-06-21",
     "status_legend": {
       "open": "Public download, public annotations, public code, or application-based access is clearly documented.",
       "request": "Public project exists, but dataset access requires license, form, approval, or institutional agreement.",
@@ -15,22 +15,22 @@
     "huggingface_url": "https://huggingface.co/datasets/cy0307/awesome-egocentric-atlas"
   },
   "summary": {
-    "total_resources": 460,
-    "egocentric_resources": 456,
-    "adjacent_resources": 4,
+    "total_resources": 479,
+    "egocentric_resources": 476,
+    "adjacent_resources": 3,
     "kind_counts": {
-      "benchmark": 81,
+      "benchmark": 90,
       "collection": 1,
-      "dataset": 125,
-      "model": 226,
+      "dataset": 127,
+      "model": 235,
       "toolkit": 23
     },
     "status_counts": {
       "benchmark": 7,
-      "open": 136,
+      "open": 137,
       "partial": 12,
       "request": 4,
-      "watch": 297
+      "watch": 316
     }
   },
   "lanes": [
@@ -43,7 +43,7 @@
         "video-language",
         "generation-and-world-models"
       ],
-      "count": 118
+      "count": 124
     },
     {
       "id": "procedure-action",
@@ -53,7 +53,7 @@
         "action-and-procedure",
         "skills-and-quality"
       ],
-      "count": 134
+      "count": 139
     },
     {
       "id": "hands-3d",
@@ -66,7 +66,7 @@
         "detection-segmentation",
         "three-d-and-scene"
       ],
-      "count": 200
+      "count": 204
     },
     {
       "id": "memory-reasoning",
@@ -78,7 +78,7 @@
         "reasoning-intent-planning",
         "grounding-localization"
       ],
-      "count": 142
+      "count": 152
     },
     {
       "id": "robotics-vla",
@@ -87,7 +87,7 @@
       "families": [
         "robotics-and-vla"
       ],
-      "count": 68
+      "count": 76
     },
     {
       "id": "ar-wearables",
@@ -98,7 +98,7 @@
         "audio-and-social",
         "assistance-and-agents"
       ],
-      "count": 134
+      "count": 142
     }
   ],
   "milestones": [
@@ -108,6 +108,7 @@
       "url": "http://kitchen.cs.cmu.edu/",
       "date": "2009-06",
       "note": "The earliest egocentric dataset; launched egocentric activity recognition at the first IEEE Workshop on Egocentric Vision (CVPR 2009).",
+      "origin": "Carnegie Mellon University",
       "image": "assets/milestones/cmu-mmac.png"
     },
     {
@@ -116,6 +117,7 @@
       "url": "https://cbs.ic.gatech.edu/fpv/",
       "date": "2011-06",
       "note": "Foundational hand-object and gaze egocentric activity datasets (GTEA, CVPR 2011) that seeded egocentric action and attention research.",
+      "origin": "Georgia Tech",
       "image": "assets/milestones/gtea-gaze.png"
     },
     {
@@ -124,6 +126,7 @@
       "url": "https://www.csc.kth.se/cvap/actions/",
       "date": "2012-06",
       "note": "Classic unscripted daily-life egocentric activity corpus with object and hand annotations; the reference point for ADL-scale egocentric recognition.",
+      "origin": "UC Irvine",
       "image": "assets/milestones/adl-dataset.png"
     },
     {
@@ -132,6 +135,7 @@
       "url": "http://vision.soic.indiana.edu/projects/egohands/",
       "date": "2015-12",
       "note": "The Google Glass hand-segmentation benchmark that made hands a first-class signal for egocentric perception.",
+      "origin": "Indiana University",
       "image": "assets/milestones/egohands.png"
     },
     {
@@ -140,6 +144,7 @@
       "url": "https://epic-kitchens.github.io/",
       "date": "2020-06",
       "note": "The defining large-scale egocentric action-recognition benchmark and annual challenge suite.",
+      "origin": "Univ. Bristol / Univ. Catania",
       "image": "assets/milestones/epic-kitchens-100.png"
     },
     {
@@ -148,6 +153,7 @@
       "url": "https://ego4d-data.org/",
       "date": "2021-10",
       "note": "The 3,670-hour massive-scale benchmark suite that catalyzed the modern egocentric era.",
+      "origin": "Meta AI / Ego4D consortium",
       "image": "assets/milestones/ego4d.png"
     },
     {
@@ -156,6 +162,7 @@
       "url": "https://hoi4d.github.io/",
       "date": "2022-03",
       "note": "Large RGB-D 4D hand-object interaction dataset that moved egocentric HOI toward geometry, pose, and temporal scene understanding.",
+      "origin": "Tsinghua / Peking / Qi Zhi",
       "image": "assets/milestones/hoi4d.png"
     },
     {
@@ -164,6 +171,7 @@
       "url": "https://github.com/showlab/EgoVLP",
       "date": "2022-06",
       "note": "First egocentric video-language pretraining (EgoClip, EgoNCE) and a basis for ego representation learning.",
+      "origin": "ShowLab",
       "image": "assets/milestones/egovlp.png"
     },
     {
@@ -172,6 +180,7 @@
       "url": "http://egoschema.github.io/",
       "date": "2023-08",
       "note": "The benchmark that exposed how far models are from long-form egocentric video reasoning.",
+      "origin": "UC Berkeley",
       "image": "assets/milestones/egoschema.png"
     },
     {
@@ -180,6 +189,7 @@
       "url": "https://www.projectaria.com/datasets/",
       "date": "2023-08",
       "note": "Meta's research smart-glasses platform that opened the modern wave of AR and wearable egocentric data.",
+      "origin": "Meta Reality Labs",
       "image": "assets/milestones/project-aria.png"
     },
     {
@@ -188,6 +198,7 @@
       "url": "https://ego-exo4d-data.org/",
       "date": "2023-11",
       "note": "Synchronized ego and exo skilled-activity capture at scale; the reference for cross-view egocentric learning.",
+      "origin": "Meta AI / Ego-Exo4D consortium",
       "image": "assets/milestones/ego-exo4d.png"
     },
     {
@@ -196,6 +207,7 @@
       "url": "https://umi-gripper.github.io/",
       "date": "2024-02",
       "note": "Handheld/wrist-view manipulation interface that made robot-free in-the-wild demonstrations practical for cross-embodiment policy learning.",
+      "origin": "Stanford",
       "image": "assets/milestones/umi.png"
     },
     {
@@ -204,6 +216,7 @@
       "url": "https://facebookresearch.github.io/hot3d/",
       "date": "2024-06",
       "note": "Reference benchmark for 3D hand-object tracking from AR glasses (Project Aria and Quest 3).",
+      "origin": "Meta / Project Aria",
       "image": "assets/milestones/hot3d.png"
     },
     {
@@ -212,6 +225,7 @@
       "url": "https://arxiv.org/abs/2503.03803",
       "date": "2025-03",
       "note": "Week-long Meta Aria daily-life corpus that pushed egocentric research toward personalized memory and life-assistant reasoning.",
+      "origin": "S-Lab, NTU / LMMs-Lab",
       "image": "assets/milestones/egolife.png"
     },
     {
@@ -220,6 +234,7 @@
       "url": "https://rchalyang.github.io/EgoVLA/",
       "date": "2025-07",
       "note": "Showed vision-language-action policies can be learned from egocentric human video and transferred to robots.",
+      "origin": "UCSD / UIUC / MIT / NVIDIA",
       "image": "assets/milestones/egovla.png"
     },
     {
@@ -228,6 +243,7 @@
       "url": "https://github.com/NVIDIA/DreamDojo",
       "date": "2026-02",
       "note": "44K-hour egocentric-video robot world model with latent actions for real-time planning.",
+      "origin": "NVIDIA",
       "image": "assets/milestones/dreamdojo.png"
     },
     {
@@ -236,6 +252,7 @@
       "url": "https://arxiv.org/abs/2602.16710",
       "date": "2026-02",
       "note": "Revealed the log-linear data-scaling law for egocentric human-video VLA pretraining.",
+      "origin": "NVIDIA / UC Berkeley / UMD",
       "image": "assets/milestones/egoscale.png"
     },
     {
@@ -244,192 +261,193 @@
       "url": "https://huggingface.co/datasets/ropedia-ai/xperience-10m",
       "date": "2026-03",
       "note": "Petascale egocentric world-model corpus (10M experiences, ~1 PB) pushing egocentric data to internet scale for embodied AI and robot learning.",
+      "origin": "Ropedia",
       "image": "assets/milestones/xperience-10m.png"
     }
   ],
   "milestone_layout": {
-    "width": 1280,
-    "height": 2322,
+    "width": 1600,
+    "height": 3605,
     "cells": [
       {
         "name": "CMU-MMAC",
         "date": "2009-06",
         "kind": "dataset",
         "url": "http://kitchen.cs.cmu.edu/",
-        "x": 266.0,
+        "x": 330.0,
         "y": 266,
-        "width": 228,
-        "height": 295
+        "width": 291,
+        "height": 437
       },
       {
         "name": "GTEA / GTEA Gaze / EGTEA Gaze+",
         "date": "2011-06",
         "kind": "dataset",
         "url": "https://cbs.ic.gatech.edu/fpv/",
-        "x": 508.0,
+        "x": 639.0,
         "y": 266,
-        "width": 228,
-        "height": 295
+        "width": 291,
+        "height": 437
       },
       {
         "name": "ADL Dataset",
         "date": "2012-06",
         "kind": "dataset",
         "url": "https://www.csc.kth.se/cvap/actions/",
-        "x": 750.0,
+        "x": 948.0,
         "y": 266,
-        "width": 228,
-        "height": 295
+        "width": 291,
+        "height": 437
       },
       {
         "name": "EgoHands",
         "date": "2015-12",
         "kind": "dataset",
         "url": "http://vision.soic.indiana.edu/projects/egohands/",
-        "x": 992.0,
+        "x": 1257.0,
         "y": 266,
-        "width": 228,
-        "height": 295
+        "width": 291,
+        "height": 437
       },
       {
         "name": "EPIC-KITCHENS-100",
         "date": "2020-06",
         "kind": "dataset",
         "url": "https://epic-kitchens.github.io/",
-        "x": 266.0,
-        "y": 691,
-        "width": 228,
-        "height": 295
+        "x": 330.0,
+        "y": 853,
+        "width": 291,
+        "height": 437
       },
       {
         "name": "Ego4D",
         "date": "2021-10",
         "kind": "dataset",
         "url": "https://ego4d-data.org/",
-        "x": 508.0,
-        "y": 691,
-        "width": 228,
-        "height": 295
+        "x": 639.0,
+        "y": 853,
+        "width": 291,
+        "height": 437
       },
       {
         "name": "HOI4D",
         "date": "2022-03",
         "kind": "dataset",
         "url": "https://hoi4d.github.io/",
-        "x": 750.0,
-        "y": 691,
-        "width": 228,
-        "height": 295
+        "x": 948.0,
+        "y": 853,
+        "width": 291,
+        "height": 437
       },
       {
         "name": "EgoVLP",
         "date": "2022-06",
         "kind": "model",
         "url": "https://github.com/showlab/EgoVLP",
-        "x": 992.0,
-        "y": 691,
-        "width": 228,
-        "height": 295
+        "x": 1257.0,
+        "y": 853,
+        "width": 291,
+        "height": 437
       },
       {
         "name": "EgoSchema",
         "date": "2023-08",
         "kind": "benchmark",
         "url": "http://egoschema.github.io/",
-        "x": 267.5,
-        "y": 1116,
-        "width": 179,
-        "height": 265
+        "x": 471.0,
+        "y": 1440,
+        "width": 300,
+        "height": 442
       },
       {
         "name": "Project Aria Datasets",
         "date": "2023-08",
         "kind": "collection",
         "url": "https://www.projectaria.com/datasets/",
-        "x": 460.5,
-        "y": 1116,
-        "width": 179,
-        "height": 265
+        "x": 789.0,
+        "y": 1440,
+        "width": 300,
+        "height": 442
       },
       {
         "name": "Ego-Exo4D",
         "date": "2023-11",
         "kind": "dataset",
         "url": "https://ego-exo4d-data.org/",
-        "x": 653.5,
-        "y": 1116,
-        "width": 179,
-        "height": 265
+        "x": 1107.0,
+        "y": 1440,
+        "width": 300,
+        "height": 442
       },
       {
         "name": "Universal Manipulation Interface / UMI",
         "date": "2024-02",
         "kind": "toolkit",
         "url": "https://umi-gripper.github.io/",
-        "x": 846.5,
-        "y": 1116,
-        "width": 179,
-        "height": 265
+        "x": 630.0,
+        "y": 1902,
+        "width": 300,
+        "height": 442
       },
       {
         "name": "HOT3D",
         "date": "2024-06",
         "kind": "dataset",
         "url": "https://facebookresearch.github.io/hot3d/",
-        "x": 1039.5,
-        "y": 1116,
-        "width": 179,
-        "height": 265
+        "x": 948.0,
+        "y": 1902,
+        "width": 300,
+        "height": 442
       },
       {
         "name": "EgoLife",
         "date": "2025-03",
         "kind": "dataset",
         "url": "https://arxiv.org/abs/2503.03803",
-        "x": 450.0,
-        "y": 1511,
-        "width": 286,
-        "height": 312
+        "x": 630.0,
+        "y": 2494,
+        "width": 300,
+        "height": 442
       },
       {
         "name": "EgoVLA",
         "date": "2025-07",
         "kind": "model",
         "url": "https://rchalyang.github.io/EgoVLA/",
-        "x": 750.0,
-        "y": 1511,
-        "width": 286,
-        "height": 312
+        "x": 948.0,
+        "y": 2494,
+        "width": 300,
+        "height": 442
       },
       {
         "name": "DreamDojo",
         "date": "2026-02",
         "kind": "model",
         "url": "https://github.com/NVIDIA/DreamDojo",
-        "x": 300.0,
-        "y": 1953,
-        "width": 286,
-        "height": 312
+        "x": 471.0,
+        "y": 3086,
+        "width": 300,
+        "height": 442
       },
       {
         "name": "EgoScale",
         "date": "2026-02",
         "kind": "dataset",
         "url": "https://arxiv.org/abs/2602.16710",
-        "x": 600.0,
-        "y": 1953,
-        "width": 286,
-        "height": 312
+        "x": 789.0,
+        "y": 3086,
+        "width": 300,
+        "height": 442
       },
       {
         "name": "Xperience-10M",
         "date": "2026-03",
         "kind": "dataset",
         "url": "https://huggingface.co/datasets/ropedia-ai/xperience-10m",
-        "x": 900.0,
-        "y": 1953,
-        "width": 286,
-        "height": 312
+        "x": 1107.0,
+        "y": 3086,
+        "width": 300,
+        "height": 442
       }
     ]
   },
@@ -499,6 +517,554 @@
       "license": "cc-by-nc-4.0",
       "license_url": "https://huggingface.co/datasets/ropedia-ai/xperience-10m-sample"
     },
+    {
+      "name": "HALOMI",
+      "kind": "model",
+      "released": "2026-06",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2606.18772",
+      "paper": "https://arxiv.org/abs/2606.18772",
+      "scale": "Humanoid loco-manipulation from human demonstrations, extending UMI-style collection with egocentric head/wrist observations and head-hand trajectories",
+      "tasks": [
+        "humanoid-control",
+        "manipulation",
+        "imitation-learning",
+        "active-perception",
+        "vla"
+      ],
+      "task_families": [
+        "robotics-and-vla"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "wrist-camera-video",
+        "human-demonstrations",
+        "robot-actions",
+        "humanoid-motion"
+      ]
+    },
+    {
+      "name": "HumanoidArena",
+      "kind": "benchmark",
+      "released": "2026-06",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2606.17833",
+      "paper": "https://arxiv.org/abs/2606.17833",
+      "scale": "Benchmark for egocentric hierarchical whole-body learning with seven leg-critical humanoid-object and humanoid-scene interaction tasks",
+      "tasks": [
+        "benchmark",
+        "humanoid-control",
+        "embodied-ai",
+        "manipulation",
+        "vla"
+      ],
+      "task_families": [
+        "evaluation-and-tooling",
+        "robotics-and-vla",
+        "foundation-and-representation"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "text-instructions",
+        "robot-actions",
+        "humanoid-motion"
+      ]
+    },
+    {
+      "name": "VLESA",
+      "kind": "benchmark",
+      "released": "2026-06",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://github.com/HanjiangHu/VLESA",
+      "paper": "https://arxiv.org/abs/2606.03954",
+      "code": "https://github.com/HanjiangHu/VLESA",
+      "scale": "Egocentric-video safety-assistance benchmark with goal-conditioned safety annotations for human activity monitoring",
+      "tasks": [
+        "safety",
+        "activity-recognition",
+        "video-language",
+        "situated-assistance"
+      ],
+      "task_families": [
+        "reasoning-intent-planning",
+        "action-and-procedure",
+        "video-language",
+        "assistance-and-agents"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "annotations",
+        "text"
+      ]
+    },
+    {
+      "name": "OVO-S-Bench",
+      "kind": "benchmark",
+      "released": "2026-06",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2606.03890",
+      "paper": "https://arxiv.org/abs/2606.03890",
+      "scale": "1,680 spatial-reasoning questions over 348 continuous egocentric videos with query timestamps and evidence intervals",
+      "tasks": [
+        "spatial-reasoning",
+        "egocentric-video-qa",
+        "streaming-video-understanding",
+        "spatial-localization"
+      ],
+      "task_families": [
+        "reasoning-intent-planning",
+        "question-answering",
+        "evaluation-and-tooling",
+        "grounding-localization"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "text",
+        "temporal-grounding",
+        "annotations"
+      ]
+    },
+    {
+      "name": "EgoPro-Bench",
+      "kind": "benchmark",
+      "released": "2026-05",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2605.07299",
+      "paper": "https://arxiv.org/abs/2605.07299",
+      "scale": "Personalized proactive-interaction benchmark with 2,400 evaluation videos, 12K+ training videos, and 12 domains",
+      "tasks": [
+        "proactive-assistance",
+        "personalized-qa",
+        "streaming-video-understanding",
+        "intent-understanding"
+      ],
+      "task_families": [
+        "assistance-and-agents",
+        "question-answering",
+        "evaluation-and-tooling",
+        "reasoning-intent-planning"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "text",
+        "annotations"
+      ]
+    },
+    {
+      "name": "Pro2Assist",
+      "kind": "model",
+      "released": "2026-05",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2605.04227",
+      "paper": "https://arxiv.org/abs/2605.04227",
+      "scale": "Continuous step-aware proactive assistance with multimodal egocentric perception, including public-source data and an AR-glasses testbed",
+      "tasks": [
+        "proactive-assistance",
+        "step-grounding",
+        "real-time-assistance",
+        "procedure-understanding"
+      ],
+      "task_families": [
+        "assistance-and-agents",
+        "action-and-procedure"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "smart-glasses",
+        "text",
+        "action-labels"
+      ]
+    },
+    {
+      "name": "EgoSPT / SPOT",
+      "kind": "dataset",
+      "released": "2026-05",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2605.20085",
+      "paper": "https://arxiv.org/abs/2605.20085",
+      "scale": "Egocentric spatially prompted manipulation trajectories with first-frame object/target grounding and 3D end-effector motion",
+      "tasks": [
+        "trajectory-prediction",
+        "visual-grounding",
+        "manipulation",
+        "robot-learning",
+        "vla"
+      ],
+      "task_families": [
+        "ar-sensing-navigation",
+        "grounding-localization",
+        "robotics-and-vla"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "trajectories",
+        "object-masks",
+        "3d-pose",
+        "text-instructions"
+      ]
+    },
+    {
+      "name": "EgoBabyVLM",
+      "kind": "benchmark",
+      "released": "2026-05",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2605.19130",
+      "paper": "https://arxiv.org/abs/2605.19130",
+      "scale": "Benchmarking cross-modal learning from naturalistic infant and adult egocentric video, including Machine-DevBench and an EgoBabyVLM challenge",
+      "tasks": [
+        "child-view-learning",
+        "video-language",
+        "representation-learning",
+        "vlm-evaluation"
+      ],
+      "task_families": [
+        "foundation-and-representation",
+        "video-language",
+        "evaluation-and-tooling"
+      ],
+      "modalities": [
+        "child-view-video",
+        "egocentric-video",
+        "audio",
+        "language"
+      ]
+    },
+    {
+      "name": "EgoInteract",
+      "kind": "dataset",
+      "released": "2026-05",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2605.18214",
+      "paper": "https://arxiv.org/abs/2605.18214",
+      "scale": "Controllable egocentric-video simulator and synthetic dataset with dense spatial and temporal annotations",
+      "tasks": [
+        "egocentric-simulation",
+        "temporal-segmentation",
+        "anticipation",
+        "next-active-object",
+        "ehoi"
+      ],
+      "task_families": [
+        "generation-and-world-models",
+        "action-and-procedure"
+      ],
+      "modalities": [
+        "synthetic-egocentric",
+        "egocentric-video",
+        "annotations",
+        "scene-graphs"
+      ]
+    },
+    {
+      "name": "EggHand",
+      "kind": "model",
+      "released": "2026-05",
+      "venue": "CVPR 2026 Findings",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://jyoun9.github.io/EggHand/",
+      "paper": "https://arxiv.org/abs/2605.07642",
+      "scale": "Multimodal foundation model for egocentric hand-pose forecasting over EgoExo4D-style video-language and action signals",
+      "tasks": [
+        "hand-forecasting",
+        "3d-hand-pose",
+        "video-language",
+        "vla"
+      ],
+      "task_families": [
+        "hand-object-interaction",
+        "pose-and-body",
+        "video-language",
+        "robotics-and-vla"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "hand-pose",
+        "language",
+        "action-labels"
+      ]
+    },
+    {
+      "name": "Map-Mono-Ego",
+      "kind": "model",
+      "released": "2026-05",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2605.20889",
+      "paper": "https://arxiv.org/abs/2605.20889",
+      "scale": "Map-grounded global human-pose estimation from monocular egocentric video, with AIST-Living paired ego video and scanned environments",
+      "tasks": [
+        "3d-human-pose",
+        "scene-understanding",
+        "localization",
+        "pose-and-body"
+      ],
+      "task_families": [
+        "pose-and-body",
+        "three-d-and-scene",
+        "grounding-localization"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "human-pose",
+        "scene-scans",
+        "camera-trajectory"
+      ]
+    },
+    {
+      "name": "Being-H0.7",
+      "kind": "model",
+      "released": "2026-05",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2605.00078",
+      "paper": "https://arxiv.org/abs/2605.00078",
+      "scale": "Latent world-action model learned from egocentric videos for future-aware reasoning and VLA policy learning",
+      "tasks": [
+        "world-modeling",
+        "vla",
+        "robot-learning",
+        "future-state-prediction"
+      ],
+      "task_families": [
+        "generation-and-world-models",
+        "robotics-and-vla",
+        "reasoning-intent-planning"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "latent-actions",
+        "robot-actions"
+      ]
+    },
+    {
+      "name": "GazeVLA",
+      "kind": "model",
+      "released": "2026-04",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://gazevla.github.io/",
+      "paper": "https://arxiv.org/abs/2604.22615",
+      "scale": "Vision-language-action policy that pretrains on large-scale egocentric human data to capture gaze, intention, and action before robot fine-tuning",
+      "tasks": [
+        "vla",
+        "gaze",
+        "intent-understanding",
+        "manipulation",
+        "robot-learning"
+      ],
+      "task_families": [
+        "robotics-and-vla",
+        "ar-sensing-navigation",
+        "reasoning-intent-planning"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "gaze",
+        "language",
+        "robot-actions"
+      ]
+    },
+    {
+      "name": "WARPED",
+      "kind": "model",
+      "released": "2026-04",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2604.10809",
+      "paper": "https://arxiv.org/abs/2604.10809",
+      "scale": "Wrist-aligned rendering pipeline that converts monocular egocentric human demonstrations into robot policy observations with 3D Gaussian Splatting",
+      "tasks": [
+        "imitation-learning",
+        "cross-embodiment-transfer",
+        "robot-learning",
+        "manipulation"
+      ],
+      "task_families": [
+        "robotics-and-vla"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "wrist-camera-video",
+        "3d-gaussian-splatting",
+        "hand-trajectories"
+      ]
+    },
+    {
+      "name": "PIE-V",
+      "kind": "benchmark",
+      "released": "2026-04",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2604.15134",
+      "paper": "https://arxiv.org/abs/2604.15134",
+      "scale": "Mistake-aware procedural egocentric-video benchmark injecting human-plausible mistakes and recovery corrections across Ego-Exo4D scenarios",
+      "tasks": [
+        "mistake-detection",
+        "procedural-reasoning",
+        "action-and-procedure",
+        "procedural-assistance"
+      ],
+      "task_families": [
+        "action-and-procedure",
+        "reasoning-intent-planning",
+        "assistance-and-agents"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "annotations",
+        "structured-reasoning"
+      ]
+    },
+    {
+      "name": "DP-DeGauss",
+      "kind": "model",
+      "released": "2026-04",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2604.07986",
+      "paper": "https://arxiv.org/abs/2604.07986",
+      "scale": "Dynamic probabilistic Gaussian decomposition for egocentric 4D scene reconstruction, separating background, hands, and objects in first-person interaction",
+      "tasks": [
+        "4d-reconstruction",
+        "scene-reconstruction",
+        "hand-object-interaction",
+        "neural-scene-reconstruction"
+      ],
+      "task_families": [
+        "three-d-and-scene",
+        "hand-object-interaction"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "3d-gaussian-splatting",
+        "hand-object-motion",
+        "4d-scene"
+      ]
+    },
+    {
+      "name": "LifeEval",
+      "kind": "benchmark",
+      "released": "2026-03",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2603.00490",
+      "paper": "https://arxiv.org/abs/2603.00490",
+      "scale": "4,075 high-quality QA pairs over continuous first-person streams for real-time task-oriented human-AI collaboration in daily life",
+      "tasks": [
+        "egocentric-video-qa",
+        "interactive-assistance",
+        "real-time-assistance",
+        "multimodal-reasoning"
+      ],
+      "task_families": [
+        "question-answering",
+        "assistance-and-agents",
+        "reasoning-intent-planning"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "dialogue",
+        "text"
+      ]
+    },
+    {
+      "name": "SAVA-X",
+      "kind": "benchmark",
+      "released": "2026-03",
+      "venue": "CVPR 2026",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2603.12764",
+      "paper": "https://arxiv.org/abs/2603.12764",
+      "scale": "Ego-to-exo imitation-error detection over asynchronous, length-mismatched egocentric and exocentric videos, evaluated with EgoMe",
+      "tasks": [
+        "mistake-detection",
+        "cross-view",
+        "imitation-learning",
+        "procedural-assistance"
+      ],
+      "task_families": [
+        "action-and-procedure",
+        "evaluation-and-tooling",
+        "robotics-and-vla",
+        "assistance-and-agents"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "exocentric-video",
+        "annotations"
+      ]
+    },
+    {
+      "name": "AG-EgoPose",
+      "kind": "model",
+      "released": "2026-03",
+      "venue": "arXiv",
+      "year": 2026,
+      "status": "watch",
+      "scope": "egocentric",
+      "url": "https://arxiv.org/abs/2603.25175",
+      "paper": "https://arxiv.org/abs/2603.25175",
+      "scale": "Attention-guided egocentric 3D human-pose estimation from fisheye camera input with dual motion/spatial streams",
+      "tasks": [
+        "egocentric-3d-pose",
+        "3d-human-pose",
+        "pose-and-body",
+        "motion"
+      ],
+      "task_families": [
+        "pose-and-body"
+      ],
+      "modalities": [
+        "fisheye-video",
+        "3d-human-pose",
+        "motion"
+      ]
+    },
     {
       "name": "Ego4D",
       "kind": "dataset",
@@ -2061,21 +2627,32 @@
     {
       "name": "Minerva-Ego",
       "kind": "benchmark",
-      "released": "2025-05",
+      "released": "2026-05",
       "venue": "arXiv",
-      "year": 2025,
+      "year": 2026,
       "status": "open",
-      "scope": "adjacent",
+      "scope": "egocentric",
       "url": "https://github.com/google-deepmind/neptune",
-      "paper": "https://arxiv.org/abs/2505.00681",
-      "scale": "Part of the Neptune / MINERVA long-video reasoning collection over web (YouTube) videos with manually annotated reasoning traces",
+      "paper": "https://arxiv.org/abs/2605.15342",
+      "scale": "Complex egocentric visual-reasoning benchmark with multi-step multimodal questions, dense reasoning traces, and spatiotemporal object masks",
       "tasks": [
         "multistep-reasoning",
-        "long-video-qa",
+        "egocentric-video-qa",
+        "spatial-reasoning",
+        "visual-grounding",
         "reasoning-traces"
       ],
       "task_families": [
-        "memory-and-long-context"
+        "memory-and-long-context",
+        "question-answering",
+        "reasoning-intent-planning",
+        "grounding-localization"
+      ],
+      "modalities": [
+        "egocentric-video",
+        "text",
+        "segmentation",
+        "structured-reasoning"
       ]
     },
     {

styles.css

@@ -62,6 +62,7 @@ select:focus-visible {
 img {
   display: block;
   max-width: 100%;
+  height: auto;
 }
 
 .skip-link {
@@ -87,13 +88,13 @@ img {
   position: sticky;
   top: 0;
   z-index: 10;
-  display: flex;
+  display: grid;
+  grid-template-columns: minmax(260px, auto) minmax(0, 1fr) auto;
   align-items: center;
-  justify-content: space-between;
-  gap: 24px;
-  width: min(1360px, calc(100% - 40px));
+  gap: 18px;
+  width: min(1480px, calc(100% - 40px));
   margin: 0 auto;
-  padding: 16px 0;
+  padding: 12px 0;
   border-bottom: 1px solid rgba(204, 221, 225, 0.62);
   background: rgba(248, 251, 251, 0.86);
   backdrop-filter: blur(16px);
@@ -113,11 +114,17 @@ img {
 
 .brand {
   gap: 12px;
+  min-width: 0;
   font-weight: 760;
   text-decoration: none;
   letter-spacing: 0;
 }
 
+.brand span {
+  min-width: 0;
+  overflow-wrap: break-word;
+}
+
 .brand-logo {
   width: 34px;
   height: 34px;
@@ -127,10 +134,24 @@ img {
 }
 
 .nav-links {
-  gap: 22px;
+  justify-self: center;
+  gap: 4px;
+  min-width: 0;
+  max-width: 100%;
+  overflow-x: auto;
+  padding: 4px;
+  border: 1px solid rgba(204, 221, 225, 0.72);
+  border-radius: var(--radius);
+  background: rgba(255, 255, 255, 0.66);
   color: var(--muted);
   font-size: 14px;
   font-weight: 650;
+  scrollbar-width: none;
+  box-shadow: 0 8px 22px rgba(15, 42, 51, 0.045);
+}
+
+.nav-links::-webkit-scrollbar {
+  display: none;
 }
 
 .nav-links a,
@@ -138,11 +159,150 @@ img {
   text-decoration: none;
 }
 
+.nav-links a {
+  display: inline-flex;
+  flex: 0 0 auto;
+  min-height: 34px;
+  align-items: center;
+  padding: 0 10px;
+  border-radius: 6px;
+}
+
 .nav-links a:hover,
 .text-link:hover {
   color: var(--teal);
 }
 
+.nav-links a:hover {
+  background: #edf6f7;
+}
+
+.header-actions {
+  display: flex;
+  min-width: 0;
+  justify-content: flex-end;
+}
+
+.language-switcher {
+  position: relative;
+  min-width: 0;
+}
+
+.language-toggle {
+  display: flex;
+  gap: 10px;
+  min-width: 156px;
+  min-height: 42px;
+  align-items: center;
+  padding: 6px 12px;
+  border: 1px solid var(--line);
+  border-radius: var(--radius);
+  background: #ffffff;
+  color: var(--muted);
+  box-shadow: 0 8px 22px rgba(15, 42, 51, 0.055);
+  cursor: pointer;
+  list-style: none;
+}
+
+.language-toggle::-webkit-details-marker {
+  display: none;
+}
+
+.language-toggle::after {
+  content: "";
+  width: 7px;
+  height: 7px;
+  flex: 0 0 auto;
+  border-right: 2px solid currentColor;
+  border-bottom: 2px solid currentColor;
+  transform: translateY(-2px) rotate(45deg);
+}
+
+.language-switcher[open] .language-toggle {
+  border-color: rgba(0, 166, 178, 0.72);
+  box-shadow: 0 12px 30px rgba(15, 42, 51, 0.08);
+}
+
+.language-switcher[open] .language-toggle::after {
+  transform: translateY(2px) rotate(225deg);
+}
+
+.language-label,
+.language-current {
+  min-width: 0;
+  overflow: hidden;
+  text-overflow: ellipsis;
+  white-space: nowrap;
+}
+
+.language-label {
+  color: var(--muted);
+  font-size: 11px;
+  font-weight: 820;
+  letter-spacing: 0.08em;
+  line-height: 1;
+  text-transform: uppercase;
+}
+
+.language-current {
+  margin-left: auto;
+  color: var(--ink);
+  font-size: 14px;
+  font-weight: 820;
+  line-height: 1.15;
+  text-align: right;
+}
+
+.language-menu {
+  position: absolute;
+  top: calc(100% + 8px);
+  right: 0;
+  z-index: 30;
+  display: grid;
+  gap: 4px;
+  min-width: 220px;
+  max-width: min(320px, calc(100vw - 28px));
+  padding: 8px;
+  border: 1px solid var(--line);
+  border-radius: var(--radius);
+  background: #ffffff;
+  box-shadow: var(--shadow);
+}
+
+.language-menu a {
+  display: flex;
+  min-width: 0;
+  min-height: 36px;
+  align-items: center;
+  justify-content: space-between;
+  gap: 12px;
+  padding: 0 10px;
+  border-radius: 6px;
+  color: var(--slate);
+  font-size: 14px;
+  font-weight: 740;
+  text-decoration: none;
+}
+
+.language-menu a:hover {
+  background: #edf6f7;
+  color: var(--teal);
+}
+
+.language-menu a.active {
+  background: #e4f6f3;
+  color: #067064;
+}
+
+.language-menu a.active::after {
+  content: "";
+  width: 8px;
+  height: 8px;
+  flex: 0 0 auto;
+  border-radius: 999px;
+  background: var(--teal);
+}
+
 .hero {
   display: grid;
   grid-template-columns: minmax(360px, 0.8fr) minmax(500px, 1.2fr);
@@ -560,45 +720,183 @@ tbody tr:hover {
 }
 
 .section.milestones {
-  width: min(1680px, calc(100vw - 24px));
+  width: min(1480px, calc(100vw - 32px));
 }
 
-.milestone-poster {
-  margin: 10px 0 0;
+.section.milestones h2 {
+  line-height: 1.12;
 }
 
-.milestone-poster-frame {
-  position: relative;
+.milestone-board {
+  display: grid;
+  gap: 38px;
+  margin-top: 28px;
+}
+
+.milestone-era {
+  display: grid;
+  grid-template-columns: minmax(190px, 230px) minmax(0, 1fr);
+  gap: 30px;
+  padding-bottom: 38px;
+  border-bottom: 1px solid var(--soft);
+}
+
+.milestone-era:last-child {
+  padding-bottom: 0;
+  border-bottom: 0;
 }
 
-.milestone-poster img {
+.milestone-era-head {
+  position: sticky;
+  top: 96px;
+  align-self: start;
+  min-width: 0;
+}
+
+.milestone-era-kicker {
+  margin: 0 0 10px;
+  color: var(--teal);
+  font-size: 12px;
+  font-weight: 840;
+  letter-spacing: 0.18em;
+  text-transform: uppercase;
+}
+
+.milestone-era-range {
   display: block;
-  width: 100%;
+  color: var(--ink);
+  font-size: clamp(28px, 3vw, 40px);
+  font-weight: 840;
+  line-height: 1.12;
+}
+
+.milestone-era-title {
+  display: block;
+  margin-top: 10px;
+  color: var(--slate);
+  font-size: 17px;
+  font-weight: 820;
+  line-height: 1.2;
+}
+
+.milestone-era-copy {
+  max-width: 220px;
+  margin: 12px 0 0;
+  color: var(--muted);
+  font-size: 13px;
+  font-weight: 650;
+  line-height: 1.45;
+}
+
+.milestone-cards {
+  display: grid;
+  grid-template-columns: repeat(auto-fit, minmax(250px, 306px));
+  gap: 16px;
+  justify-content: start;
+  align-items: stretch;
+}
+
+.milestone-card {
+  display: grid;
+  grid-template-rows: auto auto auto auto 1fr;
+  min-height: 430px;
+  min-width: 0;
+  overflow: hidden;
   border: 1px solid var(--line);
   border-radius: var(--radius);
   background: var(--paper);
-  box-shadow: var(--shadow);
+  color: var(--slate);
+  text-decoration: none;
+  box-shadow: 0 10px 24px rgba(15, 42, 51, 0.055);
+  transition: border-color 160ms ease, transform 160ms ease, box-shadow 160ms ease;
 }
 
-.milestone-link-layer {
-  position: absolute;
-  inset: 0;
-  z-index: 2;
-  pointer-events: none;
+.milestone-card > * {
+  min-width: 0;
+}
+
+.milestone-card:hover,
+.milestone-card:focus-visible {
+  border-color: rgba(0, 166, 178, 0.72);
+  box-shadow: var(--shadow-soft);
+  transform: translateY(-2px);
+}
+
+.milestone-card-media {
+  position: relative;
+  display: grid;
+  place-items: center;
+  min-width: 0;
+  aspect-ratio: 16 / 10;
+  overflow: hidden;
+  padding: 12px;
+  border-bottom: 1px solid var(--soft);
+  background: linear-gradient(180deg, #f6fbfb 0%, #eef7f7 100%);
 }
 
-.milestone-cell-link {
+.milestone-card-media img {
   position: absolute;
+  inset: 12px;
+  width: calc(100% - 24px);
+  height: calc(100% - 24px);
+  max-width: none;
+  max-height: none;
+  border-radius: 6px;
+  object-fit: contain;
+}
+
+.milestone-card-meta {
+  display: flex;
+  flex-wrap: wrap;
+  gap: 7px;
+  padding: 12px 14px 0;
+}
+
+.milestone-card-date {
+  background: var(--teal);
+  color: #ffffff;
+}
+
+.milestone-card-kind {
+  background: #fff0d4;
+  color: #9a6208;
+}
+
+.milestone-card-title {
+  display: -webkit-box;
+  margin: 11px 14px 0;
+  min-height: 44px;
+  overflow: hidden;
+  -webkit-box-orient: vertical;
+  -webkit-line-clamp: 2;
+  color: var(--ink);
+  font-size: 18px;
+  font-weight: 830;
+  line-height: 1.24;
+  overflow-wrap: break-word;
+}
+
+.milestone-card-origin {
   display: block;
-  border-radius: 14px;
-  pointer-events: auto;
-  text-decoration: none;
+  margin: 7px 14px 0;
+  color: var(--teal);
+  font-size: 12px;
+  font-weight: 820;
+  line-height: 1.35;
+  overflow-wrap: break-word;
 }
 
-.milestone-cell-link:hover,
-.milestone-cell-link:focus-visible {
-  background: rgba(0, 166, 178, 0.08);
-  box-shadow: inset 0 0 0 3px rgba(0, 166, 178, 0.64), 0 8px 22px rgba(15, 42, 51, 0.12);
+.milestone-card-note {
+  display: -webkit-box;
+  margin: 8px 14px 18px;
+  overflow: hidden;
+  -webkit-box-orient: vertical;
+  -webkit-line-clamp: 4;
+  color: var(--muted);
+  font-size: 13px;
+  font-weight: 640;
+  line-height: 1.45;
+  overflow-wrap: break-word;
 }
 
 .lane-card {
@@ -710,6 +1008,7 @@ tbody tr:hover {
 }
 
 .site-footer {
+  flex-direction: column;
   justify-content: space-between;
   gap: 20px;
   width: min(1180px, calc(100% - 40px));
@@ -718,6 +1017,7 @@ tbody tr:hover {
   border-top: 1px solid var(--line);
   color: var(--muted);
   font-size: 14px;
+  text-align: center;
 }
 
 .site-footer span:first-child {
@@ -725,15 +1025,43 @@ tbody tr:hover {
   font-weight: 820;
 }
 
+.footer-meta {
+  display: flex;
+  flex-wrap: wrap;
+  gap: 0.4rem 1.2rem;
+  align-items: center;
+  justify-content: center;
+}
+
 @media (max-width: 1040px) {
   .site-header {
     position: static;
-    flex-direction: column;
-    align-items: flex-start;
+    grid-template-columns: 1fr;
+    align-items: stretch;
+    gap: 12px;
+    padding: 12px 0 14px;
   }
 
   .nav-links {
-    flex-wrap: wrap;
+    justify-self: stretch;
+    width: 100%;
+  }
+
+  .header-actions {
+    justify-content: stretch;
+  }
+
+  .language-switcher,
+  .language-toggle {
+    width: 100%;
+  }
+
+  .language-menu {
+    left: 0;
+    right: auto;
+    grid-template-columns: repeat(2, minmax(0, 1fr));
+    width: 100%;
+    max-width: none;
   }
 
   .hero,
@@ -752,6 +1080,36 @@ tbody tr:hover {
     grid-template-columns: 1fr 1fr;
   }
 
+  .milestone-era {
+    grid-template-columns: 1fr;
+    gap: 18px;
+  }
+
+  .milestone-era-head {
+    position: static;
+    display: grid;
+    grid-template-columns: auto minmax(0, 1fr);
+    column-gap: 18px;
+    align-items: end;
+  }
+
+  .milestone-era-kicker {
+    grid-column: 1 / -1;
+  }
+
+  .milestone-era-title {
+    margin-top: 0;
+  }
+
+  .milestone-era-copy {
+    grid-column: 1 / -1;
+    max-width: 760px;
+  }
+
+  .milestone-cards {
+    grid-template-columns: repeat(auto-fit, minmax(230px, 1fr));
+  }
+
   .stat-strip {
     grid-template-columns: repeat(3, minmax(0, 1fr));
   }
@@ -784,14 +1142,64 @@ tbody tr:hover {
     width: 100%;
   }
 
+  .brand {
+    align-items: flex-start;
+  }
+
+  .brand-logo {
+    margin-top: 1px;
+  }
+
+  .nav-links a {
+    padding: 0 9px;
+  }
+
+  .nav-links {
+    display: grid;
+    grid-template-columns: repeat(2, minmax(0, 1fr));
+    overflow: visible;
+  }
+
+  .nav-links a {
+    justify-content: center;
+    text-align: center;
+  }
+
+  .nav-links a:last-child {
+    grid-column: 1 / -1;
+  }
+
+  .language-menu {
+    position: static;
+    margin-top: 8px;
+    grid-template-columns: 1fr;
+  }
+
   .stat-strip,
   .filters,
   .lane-grid,
+  .milestone-era,
   .maintenance-grid {
     display: grid;
     grid-template-columns: 1fr;
   }
 
+  .milestone-cards {
+    grid-template-columns: 1fr;
+  }
+
+  .milestone-era-head {
+    display: block;
+  }
+
+  .milestone-era-title {
+    margin-top: 10px;
+  }
+
+  .milestone-era-copy {
+    max-width: none;
+  }
+
   .table-toolbar {
     align-items: flex-start;
     flex-direction: column;