SOAR: Self-Correction for Optimal Alignment and Refinement in Diffusion Models

• URL: http://arxiv.org/abs/2604.12617v2
• Date: Fri, 17 Apr 2026 10:49:33 GMT
• Title: SOAR: Self-Correction for Optimal Alignment and Refinement in Diffusion Models
• Abstract summary: Post-training pipeline for diffusion models has two stages: supervised fine-tuning (SFT) on curated data and reinforcement learning (RL) with reward models.<n>We propose SOAR (Self-Correction for Optimal Alignment and Refinement), a bias-correction post-training method that fills this gap.<n>It is on-policy, reward-free, and provides dense per-time supervision with no credit-assignment problem.
• Score: 48.335262141752715
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: The post-training pipeline for diffusion models currently has two stages: supervised fine-tuning (SFT) on curated data and reinforcement learning (RL) with reward models. A fundamental gap separates them. SFT optimizes the denoiser only on ground-truth states sampled from the forward noising process; once inference deviates from these ideal states, subsequent denoising relies on out-of-distribution generalization rather than learned correction, exhibiting the same exposure bias that afflicts autoregressive models, but accumulated along the denoising trajectory instead of the token sequence. RL can in principle address this mismatch, yet its terminal reward signal is sparse, suffers from credit-assignment difficulty, and risks reward hacking. We propose SOAR (Self-Correction for Optimal Alignment and Refinement), a bias-correction post-training method that fills this gap. Starting from a real sample, SOAR performs a single stop-gradient rollout with the current model, re-noises the resulting off-trajectory state, and supervises the model to steer back toward the original clean target. The method is on-policy, reward-free, and provides dense per-timestep supervision with no credit-assignment problem. On SD3.5-Medium, SOAR improves GenEval from 0.70 to 0.78 and OCR from 0.64 to 0.67 over SFT, while simultaneously raising all model-based preference scores. In controlled reward-specific experiments, SOAR surpasses Flow-GRPO in final metric value on both aesthetic and text-image alignment tasks, despite having no access to a reward model. Since SOAR's base loss subsumes the standard SFT objective, it can directly replace SFT as a stronger first post-training stage after pretraining, while remaining fully compatible with subsequent RL alignment.

Related papers

• The Reward Was in Your Data All Along: Correcting Flow Matching with Discriminator-Guided RL [19.556118547263928]
  Discriminator-Guided RL (DRL) trains a discriminator to separate data from base-model samples in a pretrained representation space.<n>It reduces guidance-free FID (e.g., $9.38 to 2.62$ on SiT) and semantic-space FD (e.g., $88.2 to 19.3$ on DINOv3 for SiT)<n>It also yields a better frontier between preference reward and image fidelity under subsequent preference-based post-training.
  arXiv  Detail & Related papers  (2026-06-17T15:01:33Z)
• REVES: REvision and VErification--Augmented Training for Test-Time Scaling [53.197756110943395]
  We propose a two-stage iterative framework that alternates between online data/prompt augmentation and policy optimization.<n>We observe gains of +6.5 points over the RL baseline and +4.0 points over standard multi-turn training.
  arXiv  Detail & Related papers  (2026-06-17T10:37:23Z)
• Adaptive Negative Reinforcement for LLM Reasoning:Dynamically Balancing Correction and Diversity in RLVR [6.291433683451589]
  Reinforcement learning with verifiable rewards (RLVR) has become a highly effective method for improving the reasoning abilities of Large Language Models (LLMs)<n>Recent research shows that Negative Sample Reinforcement (NSR) can match or even exceed the performance of more complex frameworks like PPO and GRPO.<n>We propose two extensions to the NSR framework: Adaptive Negative Sample Reinforcement and Confidence-Weighted Negative Reinforcement.
  arXiv  Detail & Related papers  (2026-05-08T02:13:33Z)
• MARBLE: Multi-Aspect Reward Balance for Diffusion RL [71.6241143519038]
  Reinforcement learning has become the dominant approach for aligning diffusion models with human preferences.<n>Existing practice deal with multiple rewards by training one specialist model per reward.<n>We propose MARBLE, a gradient-space optimization framework that maintains independent advantage estimators for each reward.
  arXiv  Detail & Related papers  (2026-05-07T16:20:42Z)
• Self-Adversarial One Step Generation via Condition Shifting [11.426065945265647]
  APEX is a framework for adversarial correction signals extracted endogenously from a flow model.<n>Our 0.6B model surpasses FLUX-Schnell 12B (20$times$ more parameters) in one step quality.<n>With LoRA tuning on Qwen-Image 20B, APEX reaches a GenEval score of 0.89 at NFE=1 in 6 hours, surpassing the original 50-step teacher (0.87) and providing a 15.33$times$ inference speedup.
  arXiv  Detail & Related papers  (2026-04-14T05:54:33Z)
• Diffusion Reinforcement Learning via Centered Reward Distillation [35.979608265594685]
  We present textbf Reward Distillation (CRD), a diffusion RL framework derived from KL-regularized reward models built on forward-process fine-tuning.<n>To enable reliable text-to-image fine-tuning, we introduce techniques that explicitly control distribution drift.<n>Experiments on text-to-image post-training with textttGenEval and textttOCR rewards show that competitive SOTA reward optimization results with fast convergence and reduced reward hacking preference.
  arXiv  Detail & Related papers  (2026-03-14T21:29:33Z)
• How Far Can Unsupervised RLVR Scale LLM Training? [57.44753418846446]
  Unsupervised reinforcement learning with verifiable rewards (URLVR) offers a pathway to scale LLM training beyond the supervision bottleneck.<n>Recent works leverage model intrinsic signals, showing promising early gains, yet their potential and limitations remain unclear.<n>We classify URLVR methods into intrinsic versus external based on reward sources, then establish a unified theoretical framework.
  arXiv  Detail & Related papers  (2026-03-09T17:38:11Z)
• Reinforcement Learning with Verifiable yet Noisy Rewards under Imperfect Verifiers [90.50039419576807]
  Reinforcement Learning with Verifiable Rewards (RLVR) trains policies against automated verifiers to avoid costly human labeling.<n>To reduce vulnerability to verifier hacking, many RLVR systems collapse rewards to binary $0,1$ during training.<n>This choice carries a cost: it introduces textitfalse negatives (rejecting correct answers, FNs) and textitfalse positives (accepting incorrect ones, FPs)
  arXiv  Detail & Related papers  (2025-10-01T13:56:44Z)
• Advantage Weighted Matching: Aligning RL with Pretraining in Diffusion Models [35.36024202299119]
  We introduce textbfAdvantage Weighted Matching (AWM), a policy-gradient method for diffusion.<n>AWM raises the influence of high-reward samples and suppresses low-reward ones while keeping the modeling objective identical to pretraining.<n>This unifies pretraining and RL conceptually and practically, is consistent with policy-gradient theory, reduces variance, and yields faster convergence.
  arXiv  Detail & Related papers  (2025-09-29T17:02:20Z)
• DiffusionNFT: Online Diffusion Reinforcement with Forward Process [99.94852379720153]
  Diffusion Negative-aware FineTuning (DiffusionNFT) is a new online RL paradigm that optimize diffusion models directly on the forward process via flow matching.<n>DiffusionNFT is up to $25times$ more efficient than FlowGRPO in head-to-head comparisons, while being CFG-free.
  arXiv  Detail & Related papers  (2025-09-19T16:09:33Z)
• Flow-GRPO: Training Flow Matching Models via Online RL [80.62659379624867]
  We propose Flow-GRPO, the first method to integrate online policy reinforcement learning into flow matching models.<n>Our approach uses two key strategies: (1) an ODE-to-SDE conversion that transforms a deterministic Ordinary Differential Equation into an equivalent Differential Equation (SDE) that matches the original model's marginal distribution at all timesteps; and (2) a Denoising Reduction strategy that reduces training denoising steps while retaining the original number of inference steps.
  arXiv  Detail & Related papers  (2025-05-08T17:58:45Z)
• Training Language Models to Self-Correct via Reinforcement Learning [98.35197671595343]
  Self-correction has been found to be largely ineffective in modern large language models (LLMs) We develop a multi-turn online reinforcement learning approach, SCoRe, that significantly improves an LLM's self-correction ability using entirely self-generated data. We find that SCoRe achieves state-of-the-art self-correction performance, improving the base models' self-correction by 15.6% and 9.1% respectively on MATH and HumanEval.
  arXiv  Detail & Related papers  (2024-09-19T17:16:21Z)
• Consistent Diffusion Models: Mitigating Sampling Drift by Learning to be Consistent [97.64313409741614]
  We propose to enforce a emphconsistency property which states that predictions of the model on its own generated data are consistent across time. We show that our novel training objective yields state-of-the-art results for conditional and unconditional generation in CIFAR-10 and baseline improvements in AFHQ and FFHQ.
  arXiv  Detail & Related papers  (2023-02-17T18:45:04Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.