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

• URL: http://arxiv.org/abs/2607.14522v1
• Date: Thu, 16 Jul 2026 03:19:53 GMT
• Title: A Continuous-Time Reinforcement Learning Framework for Fine-Tuning Discrete Diffusion Models
• Abstract summary: We formulate reinforcement learning (RL) in continuous time with discrete state spaces and possibly arbitrary action spaces.<n>We develop a complete continuous-time RL framework for fine-tuning score-based discrete diffusion models.
• Score: 9.133661763917374
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: 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.

Related papers

• Scalable Maximum Entropy Reinforcement Learning for Diffusion Policies via Adjoint Matching [25.3394501858201]
  Diffusion policies represent complex action distributions in reinforcement learning (RL)<n>We introduce a highly efficient algorithm for optimizing diffusion policies by leveraging recent advances in optimal control.<n>Our approach is based on adjoint matching, which enables simulation-free training and circumvents the need for explicit likelihood estimation.
  arXiv  Detail & Related papers  (2026-06-21T18:32:25Z)
• HAD: Combining Hierarchical Diffusion with Metric-Decoupled RL for End-to-End Driving [51.268878540511054]
  We propose HAD, an end-to-end planning framework with a Hierarchical Diffusion Policy.<n>We show that HAD achieves new state-of-the-art performance on both NAVSIM and HUGSIM.
  arXiv  Detail & Related papers  (2026-04-04T04:12:47Z)
• dTRPO: Trajectory Reduction in Policy Optimization of Diffusion Large Language Models [71.26391195427878]
  Diffusion Large Language Models (dLLMs) introduce a new paradigm for language generation.<n>In this work, we aim to improve the policy optimization for dLLMs by reducing the cost of the trajectory probability calculation.<n>We evaluate dTRPO on 7B dLLMs across instruction-following and reasoning benchmarks.
  arXiv  Detail & Related papers  (2026-03-19T11:55:52Z)
• Rethinking the Trust Region in LLM Reinforcement Learning [72.25890308541334]
  Proximal Policy Optimization (PPO) serves as the de facto standard algorithm for Large Language Models (LLMs)<n>We propose Divergence Proximal Policy Optimization (DPPO), which substitutes clipping with a more principled constraint.<n>DPPO achieves superior training and efficiency compared to existing methods, offering a more robust foundation for RL-based fine-tuning.
  arXiv  Detail & Related papers  (2026-02-04T18:59:04Z)
• Multi-Objective Reward and Preference Optimization: Theory and Algorithms [3.316593788543852]
  This thesis develops theoretical frameworks and algorithms that advance constrained reinforcement learning (RL) across control, preference learning, and alignment of large language models.<n>ACPO, e-COP, warmPref-PS, PSPL, and MOPO advance RL across average-cost, episodic, and preference-driven paradigms.<n> Collectively, the thesis unifies RL across average-cost, episodic, and preference-driven paradigms, delivering theoretical advances and practical tools for safe and aligned decision-making.
  arXiv  Detail & Related papers  (2025-12-11T12:51:21Z)
• Policy Regularized Distributionally Robust Markov Decision Processes with Linear Function Approximation [10.35045003737115]
  Decision-making under distribution shift is a central challenge in reinforcement learning (RL), where training and deployment environments differ.<n>We propose DR-RPO, a model-free online policy optimization method that learns robust policies with sublinear regret.<n>We show that DR-RPO can achieve suboptimality bounds and sample efficiency in robust RL, matching the performance of value-based approaches.
  arXiv  Detail & Related papers  (2025-10-16T02:56:58Z)
• Stabilizing Policy Gradients for Sample-Efficient Reinforcement Learning in LLM Reasoning [77.92320830700797]
  Reinforcement Learning has played a central role in enabling reasoning capabilities of Large Language Models.<n>We propose a tractable computational framework that tracks and leverages curvature information during policy updates.<n>The algorithm, Curvature-Aware Policy Optimization (CAPO), identifies samples that contribute to unstable updates and masks them out.
  arXiv  Detail & Related papers  (2025-10-01T12:29:32Z)
• Decision Flow Policy Optimization [53.825268058199825]
  We show that generative models can effectively model complex multi-modal action distributions and achieve superior robotic control in continuous action spaces.<n>Previous methods usually adopt the generative models as behavior models to fit state-conditioned action distributions from datasets.<n>We propose Decision Flow, a unified framework that integrates multi-modal action distribution modeling and policy optimization.
  arXiv  Detail & Related papers  (2025-05-26T03:42:20Z)
• Sequential Monte Carlo for Policy Optimization in Continuous POMDPs [10.675160254513615]
  We introduce a novel policy optimization framework for continuous partially observable Markov decision processes.<n>Our method casts policy learning as probabilistic inference in a non-Markovian Feynman--Kac model.<n>We demonstrate the effectiveness of our algorithm across standard continuous POMDP benchmarks.
  arXiv  Detail & Related papers  (2025-05-22T14:45:46Z)
• Iterative Amortized Policy Optimization [147.63129234446197]
  Policy networks are a central feature of deep reinforcement learning (RL) algorithms for continuous control. From the variational inference perspective, policy networks are a form of textitamortized optimization, optimizing network parameters rather than the policy distributions directly. We demonstrate that iterative amortized policy optimization, yields performance improvements over direct amortization on benchmark continuous control tasks.
  arXiv  Detail & Related papers  (2020-10-20T23:25:42Z)
• Formal Controller Synthesis for Continuous-Space MDPs via Model-Free Reinforcement Learning [1.0928470926399565]
  A novel reinforcement learning scheme to synthesize policies for continuous-space Markov decision processes (MDPs) is proposed. A key contribution of the paper is to leverage the classical convergence results for reinforcement learning on finite MDPs. We present a novel potential-based reward shaping technique to produce dense rewards to speed up learning.
  arXiv  Detail & Related papers  (2020-03-02T08:29:36Z)

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.