paper-with-me

Papers

DGPO: Beyond Pairwise Preferences with Directional Consistent Groupwise Optimization

2026-05-11 · Mengyi Deng, Zhiwei Li, Xin Li, Tingyu Zhu, Yulan Yuan, Zhijiang Guo, Wei Wang arxiv

Although Large Language Models (LLMs) have made remarkable progress, current preference optimization methods still struggle to align directional consistency while preserving reasoning diversity. To address this limitation, we propose Directional-Groupwise Preference Optimization (DGPO), a lightweight framework that aggregates supervision signals at the group level and explicitly models direction-aware alignment through multi-candidate comparisons. DGPO organizes forward and reverse question-answer instances into structured sets and optimizes a margin-based likelihood objective that separates coherent reasoning paths from inconsistent alternatives. This group-wise formulation captures richer relative information than pairwise objectives and reinforces consistency across diverse reasoning pathways. Empirical results show that our constructed reverse data yields a 3.2% average improvement across five benchmarks, while DGPO further delivers consistent gains across multiple datasets and model families, achieving average accuracy improvements of up to 3.6%.

📄 PDF Abstract BibTeX arXiv:2605.10863

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

DGPO: RL-Steered Graph Diffusion for Neural Architecture Generation

2026-02-22 · Aleksei Liuliakov, Luca Hermes, Barbara Hammer arxiv

Reinforcement learning fine-tuning has proven effective for steering generative diffusion models toward desired properties in image and molecular domains. Graph diffusion models have similarly been applied to combinatori…

Neural Architecture SearchReinforcement Learning

Reinforcing Diffusion Models by Direct Group Preference Optimization

2025-10-09 · Yihong Luo, Tianyang Hu, Jing Tang arxiv

While reinforcement learning methods such as Group Relative Preference Optimization (GRPO) have significantly enhanced Large Language Models, adapting them to diffusion models remains challenging. In particular, GRPO dem…

Reinforcement Learning

From $\log π$ to $π$: Taming Divergence in Soft Clipping via Bilateral Decoupled Decay of Probability Gradient Weight

2026-03-15 · Xiaoliang Fu, Jiaye Lin, Yangyi Fang, Chaowen Hu 외 arxiv

Reinforcement Learning with Verifiable Rewards (RLVR) has catalyzed a leap in Large Language Model (LLM) reasoning, yet its optimization dynamics remain fragile. Standard algorithms like GRPO enforce stability via "hard …

Reinforcement Learning

Prediction-Powered Ranking of Large Language Models

2024-02-27 · Ivi Chatzi, Eleni Straitouri, Suhas Thejaswi, Manuel Gomez Rodriguez

Large language models are often ranked according to their level of alignment with human preferences -- a model is better than other models if its outputs are more frequently preferred by humans. One of the popular ways t…

ChatbotLanguage ModellingLarge Language ModelPrediction

DGPO: Distribution Guided Policy Optimization for Fine Grained Credit Assignment

2026-05-05 · Hongbo Jin, Rongpeng Zhu, Zhongjing Du, Xu Jiang 외 arxiv

Reinforcement learning is crucial for aligning large language models to perform complex reasoning tasks. However, current algorithms such as Group Relative Policy Optimization suffer from coarse grained, sequence level c…

Reinforcement Learning