ProRL: Effective Reinforcement Learning for Proactive Recommendation via Rectified Policy Gradient Estimation Paper • 2605.28293 • Published May 27 • 88
