Arya Maran

Ai Researcher @University Of Florida Research

Working in a team of 4 to construct an end-to-end ML conference-ready paper by mid-September.– Our approach introduces a multi-stage process that prioritizes uncertainty quantification, dynamicevidence assessment, and adaptive refinement.– Instead of directly soliciting a comprehensive answer, our method first employs a series of \"tiny probes\", concise, targeted queries, to extract specific information from an LLM based on retrieved documents.