Research map / Reinforcement learning / Visual and representation RL
Extracting Reward Functions from Diffusion Models
Felipe Pinto Coelho Nuti, Tim Franzmeyer, Joao F. Henriques
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Steer the medium-replay base diffusion model with the learned relative reward | Locomotion diffusion models · Hopper, HalfCheetah, and Walker2D | performance | Improved |
| Steer the Walker2D base model with the learned relative reward | Walker2D diffusion model · Walker2D | performance relative to unsteered model: 36.17% relative improvement | Improved |
| Apply the learned relative reward to a new medium-performance diffusion model unseen during reward training | Locomotion diffusion models · Locomotion domains | performance | Improved |
| Learn a relative reward from Safe Stable Diffusion and base Stable Diffusion | Stable Diffusion image generation · I2P prompt dataset | safe-versus-base image classification accuracy: over 90% | Improved |
| Reduce the training dataset size (ablation) | Relative reward estimator · Maze2D | performance | Hurt |
Closest papers by meaning
- Score Regularized Policy Optimization through Diffusion Behavior (ICLR 2024 · similarity 0.89)
- Confronting Reward Overoptimization for Diffusion Models: A Perspective of Inductive and Primacy Biases (ICML 2024 · similarity 0.89)
- Designing Reinforcement Learning for Diffusion Models: A Unified Path-Space View (NeurIPS 2026 · similarity 0.89)
- Directly Fine-Tuning Diffusion Models on Differentiable Rewards (ICLR 2024 · similarity 0.88)
- Prior-Guided Diffusion Planning for Offline Reinforcement Learning (NeurIPS 2025 · similarity 0.88)
- Diffusion Spectral Representation for Reinforcement Learning (NeurIPS 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
