paper-with-me

Papers

DACA-GRPO: Denoising-Aware Credit Assignment for Reinforcement Learning in Diffusion Language Models

2026-05-08 · Amin Karimi Monsefi, Dominic Culver, Nikhil Bhendawade, Lokesh Boominathan, Manuel R. Ciosici, Yizhe Zhang, Irina Belousova arxiv

Diffusion large language models are a compelling alternative to autoregressive models, yet existing RL methods for diffusion treat all denoising steps as equally important and rely on biased, high-variance likelihood estimates. We identify two fundamental weaknesses: the absence of temporal credit assignment across the denoising trajectory, and the systematic bias of mean-field likelihood estimates used for policy optimization. To address these, we propose Denoising-Aware Credit Assignment for GRPO (DACA-GRPO), a lightweight, plug-and-play enhancement for any GRPO-style trainer. DACA-GRPO introduces two complementary mechanisms: Denoising Progress Scores, which extract per-token importance weights from intermediate predictions at no additional forward cost, and Stratified Masking Likelihood, which partitions token positions into strata so that each token is predicted with most of the sequence as context, reducing the mean-field bias. Applied on top of three GRPO base methods, DACA-GRPO achieves consistent improvements across seven benchmarks spanning mathematical reasoning, code generation, constraint satisfaction, and constrained generation, with gains of up to 5.6pp on math reasoning, 7.4pp on code generation, 36.3pp on constraint satisfaction, and 5.9pp on JSON schema adherence.

📄 PDF Abstract BibTeX arXiv:2605.16342

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement LearningMathematical ReasoningCode Generation

Similar Papers 제목 키워드 기반

Learning to Credit the Right Steps: Objective-aware Process Optimization for Visual Generation

2026-04-21 · Rui Li, Ke Hao, Yuanzhi Liang, Haibin Huang 외 arxiv

Reinforcement learning, particularly Group Relative Policy Optimization (GRPO), has emerged as an effective framework for post-training visual generative models with human preference signals. However, its effectiveness i…

Reinforcement LearningVideo Generation

Element-Aware Group Learning for E-Commerce Image Generation

2026-08-01 · Jingtong Chen, Jiahui Wang, Xue Zhao, ShaoGuo Liu 외 arxiv

Recent advances in image generation and editing have made prompt quality a key bottleneck for e-commerce creatives. Vision-language models (VLMs) can generate image-editing prompts from product images and metadata, but f…

Image Generation

TreeGRPO: Tree-Advantage GRPO for Online RL Post-Training of Diffusion Models

2025-12-09 · Zheng Ding, Weirui Ye arxiv

Reinforcement learning (RL) post-training is crucial for aligning generative models with human preferences, but its prohibitive computational cost remains a major barrier to widespread adoption. We introduce \textbf{Tree…

Reinforcement Learning

On the Policy Gradient Foundations of Group Relative Policy Optimization: Credit Assignment, Gradient Sparsity, and Rank Collapse

2026-06-28 · Amritansh Mishra, Supriyo Chakraborty, Berkcan Kapusuzoglu arxiv

Group Relative Policy Optimization (GRPO) eliminates the learned critic in PPO by using the mean reward of grouped rollouts as a baseline. We provide a rigorous derivation of GRPO from first principles of the policy grad…

Multi-GRPO: Multi-Group Advantage Estimation for Text-to-Image Generation with Tree-Based Trajectories and Multiple Rewards

2025-11-30 · Qiang Lyu, Zicong Chen, Chongxiao Wang, Haolin Shi 외 arxiv

Recently, Group Relative Policy Optimization (GRPO) has shown promising potential for aligning text-to-image (T2I) models, yet existing GRPO-based methods suffer from two critical limitations. (1) \textit{Shared credit a…

Text-to-Image Generation