Publications

LAMDA: Large Language Model as Decision Analyst

Yifan Hong, Sizhong Qin, Chen Wang

Published in Decision Analysis, 2026

LAMDA converts natural-language descriptions of decisions under risk into verified influence diagrams that support better decisions than standalone or chain-of-thought LLMs, while preserving human review and refinement.

State Sensitivity in an Additive Discovery Game

Yifan Hong, Chen Wang, Bonan Zhao

Published in CogSci, 2025

Human strategies deviates from the optimal policy for the additive discovery game, e.g., explore excessively.

A Rational Model of Dimension-reduced Human Categorization

Yifan Hong, Chen Wang

Published in CogSci, 2025

A rational analysis of human categorization with simultaneous learning of low-dimensional local features via PPCA.