# Safe, Efficient, and Robust Reinforcement Learning for Ranking and Diffusion Models

arXiv: https://arxiv.org/abs/2510.15429 · HTML: https://mlsanity.com/paper/2510.15429/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Diffusion and flow policies for robotics, one of 14 directions in Diffusion models.
- Builds on 4 of its 5 references that are in the map.
- First posted on arXiv 17 Oct 2025.

## Placed in

- Diffusion models / [Diffusion and flow policies for robotics](https://mlsanity.com/diffusion/diffusion-and-flow-policies-for-robotics/)
- Deep reinforcement learning methods / [RL theory and sample complexity](https://mlsanity.com/deeprl/rl-theory-and-sample-complexity/)

## Builds on (most cited first, 4 of 4)

- [A Simple and Effective Reinforcement Learning Method for Text-to-Image Diffusion Fine-tuning](https://mlsanity.com/paper/2503.00897/) (2025)
- [Optimal Baseline Corrections for Off-Policy Contextual Bandits](https://mlsanity.com/paper/2405.05736/) (2024)
- [Safe Deployment for Counterfactual Learning to Rank with Exposure-Based Risk Minimization](https://mlsanity.com/paper/2305.01522/) (2023)
- [Practical and Robust Safety Guarantees for Advanced Counterfactual Learning to Rank](https://mlsanity.com/paper/2407.19943/) (2024)

## Abstract (arXiv)

This dissertation investigates how reinforcement learning (RL) methods can be designed to be safe, sample-efficient, and robust. Framed through the unifying perspective of contextual-bandit RL, the work addresses two major application domains - ranking and recommendation, and text-to-image diffusion models. The first part of the thesis develops theory and algorithms for safe deployment in ranking systems. An exposure-based generalisation bound is derived, leading to a counterfactual risk-minimisation objective whose solution is guaranteed not to underperform the logging policy, even with sparse feedback. This guarantee is extended to doubly robust estimators, enabling safety even under adversarial or misspecified user models and offering practitioners explicit control over permissible utility loss. The second part turns to single-action bandits, where various off-policy estimators are unified within a baseline-correction framework. A closed-form optimal baseline is proposed and shown to minimise both evaluation and policy-gradient variance, thereby improving off-policy learning reliability. The final part examines the trade-offs between efficiency and effectiveness in generative RL. A systematic study of PPO and REINFORCE motivates the Leave-One-Out PPO (LOOP) algorithm, which combines multiple diffusion trajectories with a REINFORCE-style baseline inside PPO's clipped objective. LOOP achieves PPO-level sample efficiency while producing generations that align more faithfully with textual attributes.