판 이력 — GLARE: Generative Learning via Adversarial Reward Estimation For Social Dynamics Forecasting | AIChainDay