Preference Dynamic

Preference dynamics research explores how individual preferences change over time, influenced by factors like received recommendations and personal experiences. Current research focuses on modeling these dynamics using various approaches, including reinforcement learning, contextual bandits (like HyperBandit), recurrent neural networks, and game-theoretic frameworks (like Nash CoT), aiming to optimize recommendation systems and other applications while mitigating potential harms like preference manipulation. This field is significant for improving the design of personalized systems, ensuring user well-being, and advancing our understanding of human decision-making in dynamic environments. The development of more accurate and psychologically grounded models is a key objective.