Researchers have introduced EgoCS-400K, a large-scale dataset derived from Counter-Strike gameplay, designed to advance the development of interactive world models. This dataset comprises over 400,000 first-person videos and 10,000 hours of gameplay, meticulously capturing player actions, camera movements, game states, and events. EgoCS-400K aims to bridge the gap between passive video data and the complex requirements of embodied AI by providing temporally aligned video-action-language trajectories. AI
IMPACT Provides a large-scale, action-rich dataset to train world models for interactive AI agents.
RANK_REASON The cluster describes a new research dataset published on arXiv.
Read on Hugging Face Daily Papers →
AI-generated summary · Google Gemini · from 3 sources. How we write summaries →
