RL for Diffusion Models

RL: Reinforcement Learning

Follow this topic Ask about this topic

Momentum

17 papers in the last four weeks, up 467% on the four weeks before. 0.2% of all new papers.

Jul 13Week of Sep 28

Latest papers 120

Jul 16, 2026cs.LG

A Continuous-Time Reinforcement Learning Framework for Fine-Tuning Discrete Diffusion Models

We formulate reinforcement learning (RL) in continuous time with discrete state spaces and possibly arbitrary action spaces via a stochastic control approach, where the state dynamics are modeled as a controlled continuous-time Markov chain (CTMC). We consider policy optimization problems and derive the corresponding policy gradient methods, leading to continuous-time variants of proximal policy optimization (PPO) and group relative policy optimization (GRPO). As a primary application, we develop a complete continuous-time RL framework for fine-tuning score-based discrete diffusion models. The proposed framework enables reward-driven optimization without requiring differentiability on the reward signals. In contrast to the existing GRPO-based approaches that only rely on terminal rewards, our formulation allows intermediate reward or advantage signals to be incorporated throughout the denoising trajectory. Importantly, when specialized to masked diffusion models (MDMs), our framework encompasses a rich class of policy parameterizations over the vocabulary simplex with analytically tractable probability ratios, providing a unified perspective on exploration and policy optimization in MDMs. For masked diffusion large language models (dLLMs), we further propose trajectory subsampling techniques to efficiently estimate computationally prohibitive trajectory likelihoods, reducing the computational cost of computing per-position probability ratios. We showcase the effectiveness of our methods on both low-dimensional entropy-regularized optimization problems and RL post-training of dLLMs on mathematical reasoning and coding tasks.

Zikun Zhang, Jiayuan Sheng, David D. Yao +1

Department of Industrial Engineering and Operations Research, Columbia University, New York, NY 10027

Masked Diffusion Models RL for Language Model Reasoning Diffusion Model Fine-Tuning Continuous-Time RL RL for Diffusion Models Diffusion Language Models

Jul 13, 2026cs.CV

Read It Back: Pretrained MLLMs Are Zero-Shot Reward Models for Text-to-Image Generation

In this paper, we propose SpectraReward, a training-free reward function that turns pretrained MLLMs into off-the-shelf reward models for image-generation reinforcement learning. Instead of asking the MLLM to judge a generated image or answer decomposed verification questions, SpectraReward measures how well the original prompt can be recovered from the generated image through a single image-conditioned, teacher-forced forward pass. We use the average image-conditioned prompt log-likelihood as the reward, directly reusing the MLLM's pretrained image-text alignment ability without preference labels, reward-model fine-tuning. We further introduce Self-SpectraReward, a special case for unified multimodal models where the policy's own understanding branch serves as the reward model for its generation branch, forming a closed-loop self-improving framework without external reward models or external knowledge. Extensive experiments validate SpectraReward through a broad image-generation RL study covering two diffusion models, three RL algorithms, nine reward MLLM backbones from four MLLM families spanning 4B to 235B parameters, and five out-of-distribution text-to-image benchmarks. Results show that both SpectraReward and Self-SpectraReward significantly and consistently improve generation performance and outperform prior MLLM-derived reward training methods. Further analysis reveals that larger reward MLLMs are not always better, while Self-SpectraReward can match or surpass much larger external reward models, suggesting that reward-policy alignment is a key factor for effective image-generation RL. Project Page: https://huangrh99.github.io/SpectraReward/

Runhui Huang, Qihui Zhang, Zhe Liu +3

The University of Hong Kong · Peking University · ByteDance Seed

T2I Generation Multimodal Reward Modeling RL for Image Generation RL for Diffusion Models

Jul 12, 2026cs.RO

A Single Diffusion-Policy Controller for Multi-Task Block Pushing with Zero-Shot Sim-to-Real Transfer

Diffusion policies have shown promising empirical performance in representing and learning complex maneuvers for robots using behavior cloning (BC). In this paper, we explore training diffusion policies from scratch using reinforcement learning (RL) for multi-task robotic manipulation. Specifically, we aim to train a single diffusion policy for block-pushing tasks with multiple shapes. The proposed framework features a simple policy loss function, which is a reweighted evidence lower bound used in BC-based diffusion policy training and can seamlessly serve as the policy learning module in RL algorithms. To address the exploration challenges arising from the absence of demonstrations, we incorporate reverse curriculum generation and objective-centric representations. Combined with the expressiveness of diffusion policies, our design supports learning of multi-task block-pushing policies in our sparse-reward simulation setting. We further evaluate whether the trained diffusion policy transfers in zero-shot to real-world tasks under varying environmental conditions including goal positions, block shapes, block weights and surface friction, providing evidence that this pipeline can transfer to our real-world block-pushing setup under the tested variations.

Haitong Ma, Haldun Balim, Yang Hu +2

1Harvard University. · 2Georgia Institute of Technology.

Sim-to-Real Transfer Diffusion Policy Robotic Manipulation Multi-Task RL RL for Diffusion Models

Jul 9, 2026cs.LG

Reinforcing the Generation Order of Multimodal Masked Diffusion Models

Diffusion Language Models (DLMs) have recently achieved substantial progress in natural language generation tasks. Recent research demonstrates that adaptive token generation ordering can significantly improve performance in mathematical reasoning and code synthesis applications. In this work, we investigate the optimization of generation order for both text-to-image synthesis and multimodal understanding. We first establish that, unlike structured problems in language generation such as Sudoku puzzles, model logits alone are insufficient for determining optimal generation sequences in text-to-image generation and multimodal understanding. To address this challenge, we introduce a learnable control module trained via Group Relative Policy Optimization (GRPO) to determine the generation order. Our results demonstrate that learning this control block substantially improves both text-to-image alignment and multimodal understanding in DLMs. In particular, it enhances the model's ability to capture fine-grained spatial relationships in generated images while also strengthening performance on multimodal reasoning and comprehension tasks. We evaluate our framework on GenEval, an object-focused benchmark for text-to-image alignment, where it achieves 4.08% relative improvements. In addition, experiments on VLMEvalKit confirm 4.85% relative improvements in multimodal understanding, highlighting the broad effectiveness of our approach.

Yidong Ouyang, Zhe Wang, Sourav Bhabesh +1

University of California, Los Angeles · AGI Foundations for AWS

Text-Image Alignment RL for Diffusion Models Multimodal Learning Multimodal Diffusion Models Multimodal Understanding

Jul 8, 2026stat.ML

Expressivity and Statistical Trade-offs in Diffusion Policy Learning

Diffusion-based policies have recently emerged as powerful policy parameterizations for reinforcement learning, representing state-conditioned action distributions as terminal laws of diffusion processes with parameterized drifts. This terminal-law representation has shown substantial expressive flexibility in practice, enabling diffusion policies to model complex, multimodal, and highly non-Gaussian action distributions; however, it remains unclear what mathematically drives this expressivity and how to fully exploit it when the policy is learned from finite data. In this paper, we identify the drift Lipschitz budget

K

as a central quantity governing the expressivity and statistical behavior of diffusion policies. We quantify expressivity through approximation: diffusion policies with

K

-Lipschitz drifts can concentrate near optimal deterministic policies and achieve value approximation error of order

1/K

; moreover, we prove a matching lower bound under nondegenerate diffusion noise. This increased expressivity comes with a statistical cost. When the drift is parameterized by neural networks, increasing

K

improves approximation but increases statistical complexity. Balancing these two terms yields a finite-sample performance gap of order

\tilde{O}(n^{-2/(m+6)})

for generic neural-network drifts, and a sharper rate

\tilde{O}(n^{-2/(m+4)})

for one-sided dissipative drift classes, where

n

is the sample size and

m

is the dimension of the state space. Numerical experiments provide empirical evidence for the sample-dependent trade-off in

K

, supporting both theoretical regimes. Our framework also suggests a practical implementation principle: choose the diffusion budget

K

according to the available sample size, and then select a neural-network architecture with the corresponding fixed Lipschitz coefficient.

Viet Vu, Renyuan Xu, Jiacheng Zhang +1

Stanford University · Chinese University of Hong Kong · Imperial College London

Diffusion Policy RL for Diffusion Models Policy Learning

Jul 8, 2026cs.LG

Selective Timestep Weighting and Advantage-Based Replay for Sample-Efficient Diffusion RLHF

Reinforcement learning from human feedback (RLHF) has emerged as a powerful paradigm for aligning generative models with human preferences. However, applying RLHF to diffusion models remains highly feedback inefficient, as existing approaches typically require large amounts of human or reward model evaluations. This limitation reduces the practicality of diffusion RLHF in realworld settings where feedback is the primary bottleneck. In this paper, we propose two complementary strategies that substantially improve the feedback efficiency of diffusion RLHF while preserving generalization to unseen prompts. Our key observation is that reward information in diffusion trajectories is unevenly distributed: not all denoising timesteps or trajectories contribute equally to learning from a reward signal. By emphasizing informative timesteps and trajectories during optimization, we obtain more effective gradient updates. First, we introduce a per-timestep weighting scheme that reweights denoising steps during policy optimization. We theoretically connect this weighting to the optimal convergence properties of proximal policy optimization (PPO) and approximate the resulting weighting trend empirically. Second, we introduce a replay mechanism that prioritizes informative trajectories, enabling the model to reuse past samples instead of repeatedly querying new rewards. Together, these strategies significantly improve the feedback efficiency of diffusion RLHF. Under identical hyperparameter settings, our approach achieves up to a 6

\times

improvement in sample efficiency compared to widely used diffusion RLHF baselines.

Eric Zhu, Abhinav Shrivastava, Soumik Mukhopadhyay

Carnegie Mellon University · University of Maryland, College Park

Diffusion Model Alignment RL for Diffusion Models Sample-Efficient RL

Jul 4, 2026cs.CV

Self-Improving Diffusion Classifiers with Minority Preference Optimization

Prior studies have demonstrated that diffusion classifiers achieve robust zero-shot classification performance. However, their effectiveness is strongly tied to the pretraining data distribution: they perform well in majority, high-density regions of the data manifold, but are significantly less accurate in minority, low-density regions. Although prior works on minority sampling have focused on generating more minority-like images, what minority sampling fundamentally enables beyond generation remains underexplored. In this paper, we reveal a direct relationship between minority sampling in generation and the perception capability of diffusion classifiers. Specifically, we show that enhancing minority sampling broadens the coverage of underrepresented regions on the data manifold, thereby improving diffusion-based recognition. To exploit this connection, we propose \textit{Self-Improving Diffusion Classifiers with Minority Preference Optimization} (MiPO), which fine-tunes a pretrained diffusion model using minority preference rewards. Using only arbitrary caption data, MiPO generates candidate samples, rewards those that better cover minority regions, and optimizes the model with LoRA and Group Relative Policy Optimization, without additional image data, external foundation models, or external reward models. This enables stable, prompt-adaptive minority sampling and translates low-density generative coverage into improved zero-shot diffusion classification. To sum up, we show that diffusion classifier perception is biased toward majority regions, demonstrate that this bias can be alleviated through minority preference optimization, and evaluate MiPO on five standard datasets.

Hyunsoo Kim, Jungmyung Wi, Soobin Um +2

1Korea University · 2Kook Min University · 3Kyung Hee University

Diffusion Model Alignment Long-Tailed Classification Diffusion Model Fine-Tuning RL for Diffusion Models Zero-Shot Image Classification

Jul 2, 2026cs.LG

Adaptive Reparametrized Time for Score-Based Diffusion Sampling

We study timestep allocation for score-based diffusion sampling, where a learned reverse-time dynamics is discretized on a finite grid. Uniform and hand-crafted schedules are standard choices, but they rely on fixed prescriptions and can therefore be suboptimal. To address this limitation, we propose Adaptive Reparameterized Time (ART), a continuous-time control formulation that learns a time change by treating the speed of the sampling clock as the control, so that a uniform grid on the learned clock induces adaptive timesteps in the original diffusion time. Based on a leading-order Euler error surrogate, ART provides a principled objective for allocating timesteps along the sampling trajectory. To solve this deterministic control problem, we introduce ART-RL, an auxiliary randomized formulation with Gaussian policies that turns schedule learning into a continuous-time reinforcement learning problem. We prove that the randomized ART-RL formulation is equivalent to ART at the optimizer level, in the sense that its optimal Gaussian policy recovers the optimal ART time-warping rate through its mean. We further establish policy evaluation and policy improvement characterizations and derive trajectory-based moment identities that yield implementable actor--critic updates for learning the schedule. Across experiments ranging from controlled low-dimensional settings to image generation, ART-RL can be plugged into existing diffusion samplers by changing only the timestep grid, consistently improving sample quality over strong baseline schedules at matched budgets while leaving the rest of the sampling pipeline unchanged. The learned schedules also exhibit broad generalization, transferring without retraining across sampling budgets, datasets, solvers, pipelines, and representation spaces.

Yilie Huang, Wenpin Tang, Xun Yu Zhou

Department of Applied Mathematics, The Hong Kong Polytechnic University, Hong Kong. · Department of Industrial Engineering and Operations Research, Columbia University, New York, NY 10027, USA. · Department of Industrial Engineering and Operations Research and Data Science Institute, Columbia University, New York, NY 10027, USA.

Diffusion Sampling Adaptive Sampling RL for Diffusion Models Diffusion Timestep Scheduling

Jul 1, 2026cs.LG

PAPA: Online Personalized Active Preference Alignment

Diffusion models are highly effective at modeling complex data distributions, including images and text. However, in applications like personalized recommender systems, the objective often shifts to modeling specific regions of the distribution that maximize user preferences-initially unknown but gradually uncovered through interactive feedback. This can naturally be framed as a reinforcement learning problem, where the goal is to fine-tune a diffusion model to maximize a reward function based on preferences. However, the main challenge lies in learning a parameterized reward model, which typically requires large-scale preference data-something that is often not feasible in practice. In this work, we introduce Personalized Active Preference Alignment PAPA, a novel method that bypasses the requirement for a parametrized reward model by directly optimizing the diffusion model using real-time user feedback. PAPA enables feedback-efficient preference alignment, drawing inspiration from the variational inference framework. We demonstrate PAPA's effectiveness through extensive experiments and ablation studies across diverse class-conditioned and fine-grained alignment tasks. Additionally, based on theoretical insights, we propose an enhanced fine-tuning strategy, referred to as EPAPA, that requires less computational budget and accelerates the fine-tuning process, further boosting PAPA's suitability for real-world deployment. Our code is made publicly available at https://github.com/NasikNafi/papa.

Anindya Sarkar, Nasik Muhammad Nafi, Isaac Lyngaas +2

Washington University in St. Louis · Oak Ridge National Laboratory

Diffusion Model Alignment Diffusion Model Fine-Tuning RL for Diffusion Models User Preference Modeling Preference Learning

Jun 30, 2026cs.CL

SLIM-RL: Risk-Budgeted Random-Masking RL for Diffusion LLMs Without Trajectory Slicing

Reinforcement learning for diffusion large language models (dLLMs) has largely moved to trajectory-aware methods. The current state of the art, TraceRL, holds that random masking is mismatched with the model's inference trajectory, and it reconstructs that trajectory during training by slicing each rollout into up to K/s trajectory-aligned training samples, a cost that grows with the block size K. We show that this mismatch can be mitigated without reconstructing the trajectory. Our method, SLIM-RL, bounds the commit risk of each rollout step with a tau-budget decoder, reducing aggregate commit risk in the training data. During optimization, SLIM-RL trains on these risk-controlled rollouts with a trace-free random-masking objective that adapts variance-reduction tools, combining sequence-level importance sampling, deterministic quadrature over masking levels under a mean-preserving, monotonically decreasing per-block mask schedule that we introduce. On SDAR-4B, SLIM-RL matches TraceRL's best MATH500 accuracy on only 0.46x its training samples at block size 16, improving over TraceRL by 6.32% on MATH500 and 11.05% on GSM8K under matched dynamic sampling. At block size 4, the 4B SLIM-RL surpasses the larger LLaDA-8B and Dream-7B dLLMs on math, exceeding LLaDA-8B by 10.76% on MATH500 while staying below the autoregressive Qwen2.5-7B. On code, it improves over TraceRL by 4.20% on MBPP and 3.65% on HumanEval. The tau-budget decoder transfers training-free across LLaDA, Dream, and SDAR. The source code is available at https://github.com/laolaorkkkkk/SLIM-RL .

Ruikang Zhao, Zhenting Wang, Han Gao +1

Technical University of Denmark · MBZUAI Institute of Foundation Models · Iowa State University +1

Reinforcement Learning RL for Language Model Reasoning Diffusion Model Fine-Tuning RL for Diffusion Models Diffusion Language Models

Jun 29, 2026cs.LG

Diffusion Fine-tuning with Rewarded Moment Matching Distillation

Distillation and Reinforcement Learning (RL) fine-tuning are the primary pillars of diffusion post-training. While traditionally studied in isolation, the interaction between these phases remains poorly understood, and in particular how fine-tuning impacts the generative quality of distilled models. We introduce Rewarded Moment Matching Distillation (RMMD), a novel framework that simultaneously distills diffusion models and maximizes a reward function. RMMD preserves the high-fidelity ``naturalness'' characteristic of advanced distillation (such as 8-step Moment Matching) by adapting the sampling loop for on-policy training and repurposing the distillation loss as a proxy for integral KL regularization. By evaluating the FID-Reward Pareto fronts on ImageNet, we demonstrate that RMMD achieves superior trade-offs compared to single-step baselines (DI++) and multi-step competitors (DRaFT, HyperNoise). Finally, we apply RMMD to GenCast, a state-of-the-art weather forecasting model, to distill it while optimizing the Continuous Ranked Probability Score (CRPS) metric. The resulting distilled model achieves a 7.5x speedup while outperforming the teacher model on 93% of target weather variables, and being better calibrated. This proves that RMMD scales to complex, high-dimensional scientific domains.

Alexis Jacq, Guillaume Couairon, Valentin De Bortoli +3

1Google DeepMind

Diffusion Model Distillation Diffusion Model Fine-Tuning RL for Diffusion Models

Jun 25, 2026cs.CV

Qwen-Image-2.0-RL Technical Report

We present Qwen-Image-2.0-RL, a post-training pipeline that applies reinforcement learning from human feedback (RLHF) and on-policy distillation (OPD) to improve both the visual quality and instruction-following capability of the Qwen-Image-2.0 diffusion model. To provide reliable reward signals, we construct task-specific composite reward models by fine-tuning vision-language models with a pointwise scoring paradigm and chain-of-thought reasoning. For text-to-image generation, the reward models cover alignment, aesthetics, and portrait fidelity dimensions. For image editing tasks, the reward system addresses instruction-following accuracy and face identity preservation. Building on this reward system, we develop a scalable GRPO-based RL training framework, incorporating a hybrid classifier-free guidance (CFG) strategy to preserve pre-trained knowledge, prompt curation via intra-group reward range filtering, and per-category reward weight calibration. To merge the task-specialized RL policies for T2I and editing, we propose on-policy distillation as the final training stage, which consolidates multiple teachers into a single student model through trajectory-level velocity matching. Extensive evaluation shows that Qwen-Image-2.0-RL achieves 57.84 overall score on Qwen-Image-Bench (+2.61 over the base model), Elo ratings of 1193 in text-to-image arena (+78) and 1349 in image edit arena (+93), demonstrating consistent gains in aesthetic quality, prompt adherence, and editing accuracy.

Yixian Xu, Kaiyuan Gao, Yuxiang Chen +25

Equal contribution. · Corresponding author.

Reward Modeling RL for Diffusion Models Diffusion-Based Image Editing T2I Diffusion Models RL for Image Editing RL from Human Feedback

Jun 25, 2026cs.CV

SpatialFlow-GRPO: Where Spatial Credit Drives Image Editing

Recent online reinforcement learning has substantially improved image editing quality. However, existing Flow-GRPO-style methods usually rely on a single whole-image reward, which makes fine-grained editing optimization difficult. We observe that a key obstacle in image editing is this spatial uniformity assumption: a whole-image reward cannot distinguish how different spatial regions contribute to image quality. To address this issue, we propose SpatialFlow-GRPO, a training framework that introduces spatially fine-grained reward feedback. The framework converts region-aware rewards into semantic-region-level optimization signals and aligns region advantages with the corresponding latent positions during policy updates. We also train a region-aware reward model, SFReward, construct SFReward-14K with region-annotated editing samples, and introduce MultiEditBench to evaluate multi-region editing ability. On OmniGen2 and FLUX.2-klein-4B, SpatialFlow-GRPO outperforms Flow-GRPO on GEdit-Bench, ImgEdit-Bench, and MultiEditBench. The results show that SpatialFlow-GRPO converts local feedback into spatially aligned update signals and improves editing quality.

Yankai Yang, Yancheng Long, Wei Chen +7

Harbin Institute of Technology, Shenzhen · Kuaishou Technology

Reward Modeling Group Relative Policy Optimization RL for Image Generation RL for Diffusion Models Image Editing RL for Image Editing

Jun 23, 2026cs.AI

Accelerating Disaggregated RL for Visual Generative LLMs with Diffusion-Based Parallelism and Trainer-Assisted Generation

Reinforcement learning (RL) has become a dominant post-training paradigm, driving the emergence of high-performance RL systems such as veRL for autoregressive large language models (LLMs). In parallel, diffusion-oriented RL algorithms, e.g., DanceGRPO and FlowGRPO, have rapidly expanded the scope of RL from language reasoning to diffusion-based visual and flow-based generation. However, efficient RL systems for diffusion generative LLMs remain underexplored. Existing implementations, e.g., veRL-Omni, still rely on colocated execution, which simplifies synchronization but couples rollout and training resources, limits heterogeneous deployment, and constrains independent scaling. To this end, we introduce DigenRL, a disaggregated RL framework for diffusion-based generative LLMs that supports flexible resource allocation, accommodates heterogeneous GPUs, and facilitates efficient task scheduling. To maximally reduce the execution bubbles in the disaggregated architecture, we propose: 1) a generation-axis pipeline (GAP) and time-step parallelism (TSP) in the diffusion architecture to enable finer-grained pipelining between rollout and training; 2) an elastic trainer-assisted generation (TAG) approach to enable the trainer GPU resources to dynamically assist in executing rollout generations; and 3) a tightly one-step constrained asynchronous strategy to further utilize the tail bubble in the pipeline. Extensive experiments are conducted on three hardware testbeds with 16-32 GPUs using HunyuanVideo-13B, Wan2.1-14B, FLUX.1-12B, and QwenImage-20B generative models. Experimental results show that DigenRL achieves 1.56-2.10x throughput improvements over state-of-the-art diffusion RL systems, veRL-Omni and GenRL.

Sijie Wang, Zhengyu Qing, Zhiqiang Tan +6

School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen · Data Science and Analytics Thrust, The Hong Kong University of Science and Technology (Guangzhou) · Huawei Technologies Ltd.

Reinforcement Learning Distributed RL RL for Diffusion Models Diffusion Model Acceleration Diffusion Language Models Pipeline Parallelism

Jun 22, 2026cs.LG

Scheduling Thoughts: Learning the Order of Thought in Diffusion Language Models

Masked diffusion language models decode by iteratively unmasking tokens, where the unmasking order defines an "order of thought" that strongly influences generation quality yet is typically chosen heuristically. We derive a tractable upper bound on the sequential decoding mismatch, measured by the Kullback-Leibler divergence and expressed in terms of the model's pathwise log-likelihood, with tightness under sufficient model expressivity. This bound induces a dense self-aware reward over ordered trajectories, casting order selection as a principled policy optimization problem with a frozen denoiser. We instantiate this idea as Self-Aware Scheduling (SAS), which learns a lightweight order policy using Group Relative Policy Optimization and applies seamlessly to both any-order and semi-autoregressive decoding. On Sudoku with 1B MDM, SAS improves puzzle accuracy from 82.0% (best heuristic schedule) to 91.8%, and reaches 97.5% with second-stage fine-tuning along learned trajectories. On mathematical reasoning with LLaDA-8B, SAS improves pass@1 on GSM8K from 64% to 76% and on MBPP from 39.5% to 41%, consistently matching or exceeding heuristic schedules across generation lengths and block sizes. Project page: https://jimmyxu123.github.io/SAS

Jiawei Xu, Minghui Liu, Aakriti Agrawal +2

University of Maryland, College Park · University of California, Los Angeles.

Masked Diffusion Models Language Model Decoding LLM Inference Scheduling Diffusion Language Model Decoding RL for Diffusion Models Diffusion Language Models+1

Jun 21, 2026cs.LG

Scalable Maximum Entropy Reinforcement Learning for Diffusion Policies via Adjoint Matching

Diffusion policies have recently emerged as a powerful paradigm for representing complex action distributions in reinforcement learning (RL). However, their application to online RL remains limited by the challenge of scalable training in the absence of ground-truth data, where standard optimization techniques such as score matching are not directly applicable. In this work, we introduce a highly efficient algorithm for optimizing diffusion policies by leveraging recent advances in stochastic optimal control. Our approach is based on adjoint matching, which enables simulation-free training and circumvents the need for explicit likelihood estimation or costly backpropagation through the diffusion process. Furthermore, we propose several extensions that improve the robustness and stability of the method in practical settings. Empirical results demonstrate that our approach achieves competitive performance while significantly reducing computational overhead, making diffusion policies more viable for online RL scenarios.

Serge Thilges, Onur Celik, Denis Blessing +2

Autonomous Learning Robots, Karlsruhe Institute of Technology

Diffusion Policy Maximum Entropy RL RL for Diffusion Models

Jun 21, 2026cs.CV

Curvature-Adaptive Consistency Flow Matching: Autonomous Trajectory Optimization via Reinforcement Learning

Consistency distillation has significantly accelerated diffusion-model inference, but its sampling dynamics remain underexplored. We reveal an asymmetry: although Logit-Normal sampling priors work well for standard iterative generation, consistency distillation exhibits a different difficulty profile (e.g., U-shaped), with optimization bottlenecks concentrated at the boundary stages rather than intermediate steps. To address the limitations of static sampling under evolving learning demands, we propose Curvature-Adaptive Consistency Flow Matching (CACFM). By formulating distillation as a dynamic decision process, CACFM uses a lightweight reinforcement learning agent to probe Probability Flow ODE trajectories and construct an efficiency-oriented curriculum that prioritizes critical regions without manual scheduling. Combined with Flow-adapted DMD and adversarial consistency objectives, our RL-based scheduler achieves state-of-the-art results on large-scale models such as FLUX and SDXL, mitigating structural deformities and preserving high-frequency details in extreme few-step regimes.

Songtao Tian, Guhan Chen, Bohan Li +2

1Tsinghua University, Beijing, China · 2Huawei Large Model Data Technology Lab, Shenzhen, China

Curriculum RL Few-Step Diffusion Sampling Consistency Distillation RL for Diffusion Models Diffusion Model Inference Acceleration

Jun 17, 2026cs.DC

Spotlight: Synergizing Seed Exploration and Spot GPUs for DiT RL Post-Training

Reinforcement learning (RL) post-training of Diffusion Transformers (DiTs) is prohibitively expensive, requiring thousands of high-end GPUs. Existing works explore two directions to reduce cost: seed exploration improves training convergence by selecting high-contrast samples, yet adds compute to the critical path; spot GPUs offer 69--77% lower cost, yet sit idle during training because DiT rollouts finish nearly simultaneously, which prevents LLM-style pipelining of rollout with training. Spot preemptions further break Sequence Parallelism (SP) groups, fragmenting GPU topology. We present Spotlight, the first system that harvests spot GPUs for DiT RL post-training. Spotlight rests on two key insights we devise: (1)~we show that exploration can tolerate stale model weights because exploration that uses the model weights from the previous iteration preserves the relative ranking of random seeds, allowing exploration to run on idle spot GPUs during training. (2)~SP reconfiguration can reuse on-node state, reducing group recovery from minutes to sub-second launches. Built on these insights, Spotlight introduces three techniques: a bandit-based exploration planner that maximizes reward variance within the training time budget, elastic sequence parallelism that reconfigures SP groups on the fly via persistent schedulers and intra-node weight copying, and a preemption-aware pull-based request scheduler that balances load and commits in-flight state upon preemption. We implement Spotlight on the open-source RL platform ROLL and evaluate it on Qwen-Image post-training. Spotlight reaches the same target validation score

4\times

faster than baselines, reducing total cost by

1.4

-

6.4\times

while achieving superior image quality on DeepSeek-OCR and Geneval datasets with resolution

512\times512

and

1280\times1280

.

Ruiqi Lai, Dakai An, Wei Gao +6

NTU Singapore · Hong Kong University of Science and Technology · Alibaba Group

Diffusion Transformer RL for Image Generation RL for Diffusion Models RL Post-Training

Jun 16, 2026cs.AI

STAR: SpatioTemporal Adaptive Reward Allocation for Text-to-Image RL Post-Training

Existing RL post-training methods for text-to-image generation usually convert the final-image reward into a single scalar advantage and apply it with the same strength to the entire generative trajectory. However, text-to-image generation naturally has temporal and spatial structure: different denoising steps are responsible for different generation stages, and the content that truly determines text alignment often appears only in part of the image. This granularity mismatch makes it difficult for policy updates to focus on the generative components that actually affect the reward. To address this issue, we propose \textbf{SpatioTemporal Adaptive Reward (STAR) Allocation} for RL post-training of text-to-image diffusion and flow models. STAR uses text-image attention inside the generative model and starts from the core content that the user truly cares about in the prompt. It constructs spatial allocation maps that dynamically vary across denoising steps and rollouts, and allocates the same group-relative advantage to more relevant latent regions with almost no additional computational overhead. STAR then applies stronger policy updates to these regions through a spatially resolved policy objective. We use Stable Diffusion 3.5 Medium as the base model and evaluate on three tasks: GenEval, OCR text rendering, and PickScore. Experimental results show that STAR improves compositional semantic alignment, text rendering, and preference optimization without changing the external reward source, achieving

\mathbf{0.9759}

,

\mathbf{0.9757}

, and

\mathbf{23.60}

on GenEval, OCR, and PickScore, respectively.

Jinjie Shen, Wei Deng, Xian Hu +2

Diffusion Model Alignment Compositional T2I Generation RL for Diffusion Models Credit Assignment T2I Diffusion Models

Jun 13, 2026cs.LG

Trust-Region Diffusion Policies for Massively Parallel On-Policy RL

Reinforcement learning with massively parallel simulations has become a standard framework for developing robust, deployable policies; however, most existing approaches still rely on simple Gaussian policy parameterizations. Diffusion models provide a more expressive policy class and have shown strong performance on challenging control problems, yet most diffusion-based RL methods are designed for offline or off-policy training. In this work, we ask whether diffusion policies can be trained effectively in the massively parallel, on-policy regime. To this end, we introduce Trust-region Diffusion Policies (TruDi), which enables diffusion policies for on-policy RL with massively parallel simulations. This setting is particularly challenging because the data distribution changes quickly across updates, making stable training with complex policies difficult. TruDi addresses this by integrating a trust-region optimization rule to enforce a KL-divergence constraint over the entire diffusion trajectory. Empirically, we evaluate TruDi on a diverse set of 4 massively parallel RL benchmarks comprising a total of 73 tasks. Across these tasks, TruDi consistently outperforms or is on-par with strong baselines on standard tasks and achieves clear gains on more challenging humanoid control tasks, establishing a strong new baseline for massively parallel on-policy RL.

Huy Le, Onur Celik, Denis Blessing +6

Bosch Center for Artificial Intelligence · Autonomous Learning Robots, KIT · University of Texas at Austin +1

Diffusion Policy RL for Diffusion Models

Jun 13, 2026cs.LG

Temporal Difference Learning for Diffusion Models

Diffusion models are typically trained with objectives that focus on local denoising targets at individual time steps (or adjacent pairs), which do not enforce consistency between predictions along the denoising trajectory. This lack of cross-time consistency can degrade performance, especially for few-step samplers. We introduce a temporal difference (TD) objective that penalizes inconsistency of the model's multi-step progress along the denoising path. By reformulating the diffusion process as a Markov reward process and casting denoising as a policy evaluation problem in reinforcement learning, we derive a unified TD approach that applies to both discrete- and continuous-time diffusion formulations. We further propose a principled sample-based reweighting method that stabilizes training. Empirically, we show that using our TD training can significantly improve sample quality measured by FID, with stronger advantages when the number of sampling steps is small, highlighting its practical utility under low-computation-budget scenarios. We provide ablation studies to justify our design choices, including pairwise loss reweighting, regularization weight, and one-step stride. Overall, our TD approach can be a general drop-in that enforces cross-time consistency and improves generation quality across different diffusion generative models.

Qizhen Ying, Yangchen Pan, Victor Adrian Prisacariu +1

Department of Engineering Science, University of Oxford, Oxford, United Kingdom · School of Computer Science, Carleton University, Ottawa, Canada.

Few-Step Diffusion Sampling RL for Diffusion Models Diffusion Models

Jun 11, 2026cs.LG

QPILOTS: Efficient Test-Time Q-Steering for Flow Policies

Flow-matching and diffusion policies are expressive action generators, but optimizing them with temporal-difference reinforcement learning (RL) remains difficult. Effective policy extraction requires exploiting the critic's action gradient, yet directly backpropagating this signal through a multi-step denoising process can be numerically unstable. Existing methods work around this either by discarding gradient information, distilling the policy into a simpler one-step actor, or repeatedly fine-tuning the denoising policy as the critic improves. We propose QPILOTS, a method that leaves the original policy unmodified and steers the denoising process at inference time. At each denoising step, instead of evaluating the critic on the noisy intermediate action where critic predictions are unreliable, we first project that intermediate state to an estimate of the final clean action and compute the critic gradient there. We introduce two variants: QPILOTS-U uses a fast single-point approximation, while QPILOTS-M draws differentiable posterior samples via a learned auxiliary network. On a standard offline-to-online RL benchmark, QPILOTS achieves the best aggregate performance, reaching an average success rate of 90% across 50 tasks. We also apply QPILOTS to steer a large, frozen, pretrained Vision-Language Action (VLA) foundation model, outperforming or matching prior inference-time approaches across six manipulation tasks in simulation.

Yifan Ruan, Chenyang Cao, Andreas Burger +7

University of Toronto · Vector Institute · LG Electronics

Inference-Time Optimization Diffusion Policy Test-Time RL RL for Diffusion Models Policy Learning Offline-To-Online RL

Jun 11, 2026cs.LG

DiPOD: Diffusion Policy Optimization without Drifting Apart

RL post-training has become increasingly pivotal for improving diffusion policies, but existing diffusion policy-gradient methods are often unstable and cannot achieve reliable policy improvement. We identify the cause as the double-drift phenomenon: optimizing a variational surrogate can let the ELBO separate from the true log-likelihood, which then makes the resulting proxy policy gradient misaligned with the true policy gradient of expected return. We propose \textbf{DiPOD}, a diffusion policy optimization framework that maintains tight-bound behavior throughout training by interleaving self-distillation with policy-improving gradient updates. This leads to a simple and practical algorithm: augmenting each diffusion policy-gradient update with an on-policy ELBO regularizer. Across diffusion language model post-training and continuous-control diffusion policies, DiPOD substantially stabilizes training and reaches higher rewards than previous methods.

Haozhe Jiang, Haiwen Feng, Pieter Abbeel +3

UC Berkeley · NVIDIA

Policy Gradient Methods RL for Diffusion Models Diffusion Language Models

Jun 11, 2026cs.LG

A2D2: Fine-Tuning Any-Length Discrete Diffusion for Adaptive Decoding

Discrete diffusion models offer a simple and stable likelihood-based framework for sequence generation, recently extended to any-length settings via token insertion. Principled reward-guided fine-tuning for any-length discrete diffusion, however, remains largely unexplored. We introduce Fine-Tuning Any-Length Discrete Diffusion for Adaptive Decoding (A2D2), a unified framework for reward-guided fine-tuning of any-length discrete diffusion models via joint optimization of the insertion and unmasking policies together with a quality-based inference schedule. We derive the Radon-Nikodym derivative for the joint insertion-unmasking path measures, enabling theoretically guaranteed convergence to the intractable reward-tilted sequence distribution without requiring target samples. Building on this, we establish unmasking and insertion quality as tractable approaches for minimizing decoding error and introduce the Adaptive Joint Decoding (AJD) loss, which provably yields the optimal path measure that generates the reward-tilted distribution. Empirically, A2D2 improves reward optimization while enhancing generation flexibility and accuracy over prior fixed-length fine-tuning and inference-time guidance methods.

Sophia Tang, Yuchen Zhu, Molei Tao +1

Department of Computer and Information Science, University of Pennsylvania · School of Mathematics, Georgia Institute of Technology · Department of Bioengineering, University of Pennsylvania

Diffusion Model Fine-Tuning Diffusion Language Model Decoding RL for Diffusion Models Diffusion Models

Jun 11, 2026cs.CV

Efficient Reinforcement for Visual-Textual Thinking with Discrete Diffusion Model

RL-based post-training has been widely adopted to enable interleaved visual and textual reasoning in unified multimodal models capable of both text and image generation. However, most existing approaches are built upon autoregressive (AR) unified models, which require full image regeneration during visual reasoning. In this work, we demonstrate that multimodal discrete diffusion models are effective alternatives to AR models for reinforcement learning in interleaved reasoning, owing to their ability to perform efficient visual rollouts via localized visual editing rather than full image-token regeneration. This reduces rollout computation during GRPO by 26.9% compared to AR baselines, with minimal performance drop. Despite the improved efficiency, we find that joint reward assignment, which employs a shared reward signal across modalities, introduces cross-modal interference between unrelated image and text token sequences during RL updates. To address this issue, we propose factorized reward assignment, a strategy that assigns rewards independently to text and vision segments. With factorized reward assignment, our RL approach achieves an 11.2% improvement over joint reward assignment and a 38.04% improvement over the base model.

Yoonjeon Kim, Yuhta Takida, Chieh-Hsin Lai +2

KAIST · Sony AI · KAIST & AITRICS +1

Multimodal Reasoning RL for Diffusion Models RL for VLMs Multimodal Diffusion Models

Jun 7, 2026cs.CL

Back on Track: Aligning Rewards and States for Reasoning in Diffusion Large Language Models

Reinforcement learning (RL) holds immense promise for enhancing the reasoning capabilities of diffusion large language models (dLLMs). However, progress is fundamentally constrained by a dual misalignment between authentic generation trajectory and the gradient update process: (i) Process-reward misalignment. Sparse, terminal rewards are indiscriminately assigned to all intermediate steps of the generation process, failing to provide discriminative credit assignment. (ii) State-trajectory misalignment. Policy updates are often diverted toward artificial, out-of-trajectory states, squandering gradients on less informative samples. To address these limitations, we introduce Process Aligned Policy Optimization (PAPO), a novel framework that holistically aligns the RL update with the dLLM's generative trajectory via Step-Aware Process Rewards (SPR) that transform sparse terminal rewards into dense, step-wise credit, and Entropy-Guided Historical Re-enactment (EHR) that replays authentic trajectories at high-uncertainty steps. Extensive experiments on four benchmarks demonstrate that PAPO significantly outperforms baselines, achieving gains of up to 4.5% on GSM8K, 4.8% on MATH500, 42.2% on Countdown and 16.1% on Sudoku.

Yawen Shao, Jie Xiao, Kai Zhu +6

University of Science and Technology of China · Tongyi Lab · Northeastern University

RL for Language Model Reasoning Credit Assignment in RL RL for Diffusion Models Diffusion Language Models Reward Design for RL

Jun 3, 2026cs.CL

Read the Trace, Steer the Path: Trajectory-Aware Reinforcement Learning for Diffusion Language Models

Diffusion large language models (dLLMs) generate responses by iteratively unmasking and revising many positions in parallel. This process leaves a rich denoising trace depicting which tokens become confident, which remain unstable, and when commitments form. Existing dLLM reinforcement learning methods use this signal only weakly. Flat rollouts are cheap, but assign a single outcome reward to the whole trajectory. Tree rollouts provide finer, verifiable training signals by branching partial trajectories and propagating leaf rewards upward, but are compute intensive. We ask whether the denoising trace itself can provide tree-like supervision without tree-level compute. We introduce CAPR (Cached-Amortized Path Refinement), a dLLM-RL algorithm that summarizes the denoising trace into a compact path state, uses cached trajectory states to generate cheap sibling continuations, and trains a block-level value head for local block-wise supervision. Under a block-wise unmasking schedule, CAPR records path-state and block-progress features, then redistributes the final outcome reward across blocks according to the tokens revealed in each block. This trains the value head to convert one sparse reward into block-level PPO weights. CAPR therefore recovers much of the granularity of tree search while avoiding full tree expansion, reducing rollout-generation cost to roughly 0.75x that of flat rollouts and 0.6x that of tree rollouts (under standard settings). Across 4x4 Sudoku, Countdown, GSM8K, and Math500, on dense and mixture-of-experts LLaDA backbones, CAPR sets a new state of the art for RL-tuned dLLMs at 256- and 512-token budgets. On Sudoku, it matches the strongest tree-structured baseline at less than one third of the per-step compute.

Anant Khandelwal, Manish Gupta

Microsoft AI, India

RL for Language Model Reasoning Credit Assignment in RL RL for Diffusion Models Diffusion Language Models

May 31, 2026cs.LG

Fine-Tuning Diffusion Models for Molecular Generation via Reinforcement Learning and Fast Sampling

Generating molecules that simultaneously satisfy drug-like properties and conform to the 3D structure of a target protein is a core challenge in structure-based drug design (SBDD). Existing generative approaches, however, often rely on costly post-hoc processing during Sampling or require carefully curated datasets during training, yet still achieve modest gains. These limitations are especially pronounced in multi-objective settings, where balancing conflicting criteria remains a core challenge. To address these challenges, We propose FTDiff, a reinforcement learning fine-tuning framework tailored for diffusion-based molecular generation under structural constraints. To ensure stable and sample-efficient optimization, FTDiff adopts a group relative policy optimization (GRPO) style strategy. Furthermore, FTDiff builds upon a time-free pretrained diffusion model and incorporates a fast sampling mechanism that reduces the number of denoising steps, significantly accelerating both training and inference while maintaining generation quality. By optimizing a fixed threshold-aware reward, FTDiff effectively guides the model to produce valid, diverse, and high- quality molecules that balance multiple drug design objectives. Extensive experiments on benchmark datasets demonstrate that FTDiff consistently outperforms prior methods, without requiring expensive post-hoc optimization or intricate data engineering.

Guang Lin, Shikui Tu, Lei Xu

Department of Computer Science and Engineering, Shanghai Jiao Tong University, Shanghai, China · Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ), Shenzhen, Guangdong, China

Molecular Optimization Diffusion Model Fine-Tuning Structure-Based Drug Design 3D Molecular Generation RL for Diffusion Models Molecular Generation

May 31, 2026cs.LG

Lagrangian Perturbation Diffusion Steering: Latent Reinforcement Learning for Generative Policies

Behavior cloning with high-capacity generative policies achieves strong imitation performance, but is often limited by demonstration coverage and distribution shift. Direct reinforcement learning fine-tuning can improve performance, but updating large action decoders is frequently unstable and sample inefficient. We propose Lagrangian Perturbation Diffusion Steering (LP-DS), a lightweight adaptation method that improves a frozen generative policy by learning a compact noise-space perturbation before decoding. LP-DS optimizes this perturbation with a Lagrangian trust-region objective, improving downstream value while constraining deviation from the latent prior. Across RoboMimic manipulation, OpenAI Gym locomotion, and Adroit dexterous manipulation benchmarks, LP-DS improves sample efficiency, success, and return while maintaining higher action-space entropy than unconstrained noise-space steering, with return improvements of up to 25% over prior baselines. Additional evaluations with flow-matching backbones, a large vision-language-action model, and physical Franka deployment show that LP-DS is not limited to compact diffusion policies or simulated benchmarks. Project page: https://sites.google.com/view/lp-ds/home.

Hikmet Simsir, Ozgur S. Oguz

Department of Computer Engineering, Bilkent University, Ankara, Türkiye.

Robot Policy Adaptation Reinforcement Learning Constrained RL Robotic RL Diffusion Policy RL for Diffusion Models+2

May 30, 2026cs.CV

Improving Visual Representation Alignment Generation with GRPO

Recent diffusion transformers have demonstrated strong image synthesis capabilities but remain inefficient to train due to weak alignment between generative and discriminative representations. While representation alignment frameworks such as REPA improve convergence by aligning noisy denoising features with pretrained visual encoders, their externally supervised alignment loss is static and lacks adaptivity during training and inference. Existing methods rely on fixed cosine alignment or contrastive objectives, which cannot dynamically balance representation consistency and generation quality, resulting in limited discriminative benefit and failing to optimize alignment in a task-adaptive manner. To address this, we propose VRPO, a reinforcement-based optimization strategy that replaces REPA's static alignment loss with a generative representation policy optimization objective. Instead of enforcing a fixed similarity constraint, VRPO treats representation alignment as a reward-guided process: the model receives adaptive rewards based on generation fidelity, perceptual quality, and semantic coherence between the diffusion features and pretrained visual embeddings. This formulation enables the generator to continuously refine its internal representations toward semantically meaningful directions while improving image quality. Our VRPO-driven training seamlessly integrates into diffusion transformers, introducing negligible computation cost and preserving full compatibility with SiT and DiT architectures. Extensive experiments on ImageNet-256x256 demonstrate that our VRPO-Alignment substantially enhances both convergence and fidelity, achieving up to +1.8 FID improvement and 2.3x faster training compared to REPA under identical compute budgets.

Shentong Mo, Sukmin Yun

Carnegie Mellon University · Hanyang University

Diffusion Model Alignment Diffusion Transformer Diffusion-Based Image Generation RL for Diffusion Models Representation Alignment