Discover a code-free LLM-driven method for accurate, accessible eye-tracking event detection that simplifies data analysis and boosts research efficiency.
Explore MISID, a multimodal multi-turn dataset for complex intent recognition in strategic deception games, enhancing AI understanding of human behavior.
Discover how interactive program synthesis models collaborative physical tasks from narrated demos, enabling intuitive teaching and correction without codi...
Discover Gelina, a breakthrough framework that jointly synthesizes speech and gestures using interleaved tokens for natural, synchronized multimodal commun...