Fairness Begins with State: Purifying Latent Preferences for Hierarchical Reinforcement Learning in Interactive Recommendation

Researchers propose DSRM-HRL, a new framework that uses diffusion models to purify user preference data and hierarchical reinforcement learning to balance recommendation accuracy with fairness. The system addresses bias in interactive recommendation systems by separating state estimation from decision-making, achieving better outcomes on both utility and exposure equity.

• →Interactive recommendation systems suffer from popularity bias and exposure bias that distorts user preference signals.
• →The proposed DSRM-HRL framework uses diffusion models to denoise user interaction data and recover true preferences.
• →Hierarchical reinforcement learning separates long-term fairness objectives from short-term engagement optimization.
• →Experiments show the system breaks the 'rich-get-richer' feedback loop common in recommendation algorithms.
• →The approach achieves superior balance between recommendation utility and exposure equity compared to existing methods.