Recommendation Policy

Recommendation policy research focuses on designing algorithms that optimize the selection and presentation of items to users, aiming to maximize user engagement and satisfaction while mitigating potential harms. Current research emphasizes addressing biases in user feedback (e.g., herding effects), developing efficient models for sequential recommendations using continuous control and reinforcement learning, and creating more explainable and robust systems that account for user preference dynamics and adherence to recommendations. These advancements are crucial for improving the effectiveness and fairness of recommender systems across various applications, from e-commerce and entertainment to high-stakes decision-making in cyber-physical-human systems.