EPIC-KITCHENS-100
PulseAugur coverage of EPIC-KITCHENS-100 — every cluster mentioning EPIC-KITCHENS-100 across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
New TARA attack can fool AI video models with single physical texture
Researchers have developed a novel method called TARA (Trajectory-aware adversarial Alignment and Refinement Attack) to create physical adversarial textures that can mislead AI models. These textures are designed to be …
-
MacJEPA model tackles missing sensor data in egocentric videos
Researchers have introduced MacJEPA, a novel audio-visual recognition model designed to handle missing sensor data in untrimmed egocentric videos. This model addresses the challenge of temporally localized sensor outage…
-
Paper reviews language integration in video action anticipation
A new paper published on arXiv reviews the field of language-augmented video action anticipation, a task focused on predicting future human actions from partial video data. The authors propose an evidence-aware design m…
-
Audio-first triage slashes VLM calls for egocentric video captioning
Researchers have developed a novel audio-first approach for efficiently captioning long egocentric videos. This method prioritizes audio cues to decide which video segments are most relevant for analysis by a vision-lan…
-
New framework Pegasus translates human videos for robot learning
Researchers have developed Pegasus, a novel framework designed to bridge the embodiment gap in robotics. This system translates human manipulation videos into robot-learnable data by constructing a knowledge graph that …
-
New benchmarks and models advance egocentric video understanding in AI
Researchers are developing new methods and benchmarks to improve the temporal and spatial reasoning capabilities of multimodal large language models (MLLMs), particularly for egocentric video understanding. Papers intro…
-
New Embodied AI Systems Leverage Egocentric Video for Memory and Proactive Assistance
Researchers have introduced new systems and datasets for embodied AI, focusing on memory and proactive assistance from egocentric videos. MEMORA aims to equip robots with embodied action memory, using a lifecycle of for…
-
TrAction uses sparse trajectories for efficient action recognition
Researchers have developed TrAction, a novel transformer architecture for action recognition using sparse point trajectories instead of dense video. This method aims to reduce biases found in traditional models that rel…
-
TAP-JEPA model achieves second place in action anticipation challenge
Researchers have developed TAP-JEPA, a novel action anticipation model that achieved second place in the EPIC-KITCHENS-100 challenge. This model leverages frozen V-JEPA 2.1 features, utilizing a ViT-G/384 encoder and a …