Скачать или смотреть NeurIPS 2025 Research Spotlight: Conformal Information Pursuit

@PennEngineering doctoral student Kwan Ho Ryan Chan (Electrical and Systems Engineering) presents his NeurIPS 2025 research on helping LLMs (large language models) ask better follow-up questions by more effectively estimating uncertainty in interactive tasks.

This year, Penn researchers had more than 100 papers accepted to NeurIPS 2025, spanning large language models, generative AI, reinforcement learning, trustworthy AI, computational biology, neuroscience, health and the physical sciences.

Read the paper: https://arxiv.org/abs/2507.03279

This research is advised by Professors Edgar Dobriban, Hamed Hassani and René Vidal.

Explore all Penn NeurIPS 2025 papers: https://ai.seas.upenn.edu/neurips-2025/