Researchers have developed SMILE-Next, a new dataset and framework designed to improve how large language models understand and reason about laughter in real-world contexts. The system includes laughter-specific Self-Instruct for generating diverse training instructions and a Mixture-of-Laugh-Experts (MoLE) framework that uses adaptive routing to select specialized experts for different laughter-related tasks. Experiments demonstrate that this approach significantly outperforms existing multimodal LLM baselines in detecting, classifying, and reasoning about laughter. AI
IMPACT Enhances LLM capabilities in understanding nuanced human social signals, potentially improving human-AI interaction.
RANK_REASON The cluster describes a new academic paper introducing a novel dataset and framework for AI research. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
