Paper 2605.28293
ProRL: Effective Reinforcement Learning for Proactive Recommendation via Rectified Policy Gradient Estimation
- Published
- May 2026
- Research lab
- Independent
- Citations
- 0
- GitHub
- 46 stars
01 In brief
Summary
The paper introduces ProRL, a reinforcement learning framework for proactive recommender systems (PRS) that guide user preferences toward target items via intermediate recommendation paths.
The authors identify two deficiencies in standard policy gradient estimation for PRS: (1) a length shortcut, where positive-mean step-level rewards cause gradients to favor path extension over quality exploration, and (2) high gradient variance from weighting each step by the entire path reward.
ProRL addresses these with two mechanisms: Stepwise Reward Centering, which subtracts expected step rewards to neutralize length bias, and Position-Specific Advantage Estimation, which computes step-dependent baselines to reduce variance.
Experiments on MovieLens-1M, Steam, and Amazon-Book show ProRL significantly outperforms state-of-the-art baselines (including IRN, IPG, LLM-IPP, T-PRA) on CTR, Coherence, IoI, and IoR.
Cross-evaluator analysis confirms generalization beyond the training reward model.
Ablations validate each component's contribution, showing ProRL achieves stable path lengths and low gradient variance compared to alternatives like GRPO and A2C.
02 From the paper
Abstract
Proactive Recommender Systems (PRSs) aim to guide user preference shift toward target items by generating paths of intermediate recommendations. Reinforcement learning (RL) provides a principled framework for optimizing such sequential decision tasks, as path rewards can naturally capture both short-term acceptance and long-term guidance effectiveness. However, naively applying policy gradients to PRS results in deficient gradient estimation. We identify two deficiencies: (1) path-level rewards decompose into step-level rewards with positive mean, creating a length-dependent bias that causes gradients to favor path extension over meaningful exploration; (2) weighting each step by the entire path-level reward ignores the decomposition structure, leading to high gradient variance. To rectify these two deficiencies, we propose an effective RL framework ProRL with two novel mechanisms for proactive recommendation. First, Stepwise Reward Centering subtracts expected rewards to neutralize length-dependent bias, ensuring that path extension yields zero expected gradient signal. Second, Position-Specific Advantage Estimation leverages the reward decomposition structure to compute step-dependent baselines, reducing gradient variance. Together, these mechanisms yield policy gradients that precisely target path quality. Our experiments on three real-world datasets demonstrate that ProRL significantly outperforms state-of-the-art PRSs. Our code is available at https://github.com/hongruhou89/ProRL.