DeLLMa: A Framework for Decision Making Under Uncertainty with Large Language Models

Explore a groundbreaking framework for enhancing decision-making accuracy in uncertain environments using large language models (LLMs) in this hour-long talk presented by Oliver Liu from USC Information Sciences Institute. Delve into the challenges of applying LLMs to complex decision-making problems in fields like business, engineering, and medicine. Discover how the DeLLMa (Decision-making Large Language Model assistant) framework utilizes a multi-step scaffolding procedure based on decision theory and utility theory principles to significantly improve LLM performance. Learn about the validation process involving real agriculture and finance data, which demonstrated up to a 40% increase in accuracy compared to competing methods. Gain insights into the speaker's research on multimodal foundation models and their applications in algorithmic reasoning and scientific domains.

DeLLMa: A Framework for Decision Making Under Uncertainty with Large Language Models