Many strong decoding methods for text generation follow a sample-and-rerank paradigm: they draw multiple candidates, score each under a utility (reward) function using consensus across samples, and return the best one. Although effective, these methods incur high computational costs during inference due to repeated sampling and scoring. Prior attempts to amortize inference-time computation typically rely on gold references, teacher labels, or curated preference data, increasing dataset construction effort and the demand for high-fidelity reward models. We propose Consensus Group Relative Policy Optimization (C-GRPO), which distills Minimum Bayes Risk (MBR) decoding into training by formulating the consensus utility as a group-relative objective within GRPO. C-GRPO requires only a utility function and policy samples, without gold references or explicit preference labels. Under ideal conditions, we show that the objective function of C-GRPO is directionally aligned with the gradient of the expected-utility objective underlying MBR decoding, leading to a convergence guarantee. Experiments on machine translation (WMT 2024) and text summarization (XSum) demonstrate that C-GRPO successfully achieves performance comparable to MBR decoding without the associated inference-time overhead, while outperforming reference-free baseline methods.
Enforcing the consensus between Trajectory Optimization and Policy Learning for precise robot control
Inria（法国国家信息与自动化研究所）；Département d’Informatique de l’École normale supérieure, PSL Research University（巴黎文理研究大学巴黎高等师范学院计算机系）；LAAS-CNRS（法国国家科学研究中心系统分析与架构实验室）
Multi-Modal Manipulation via Multi-Modal Policy Consensus
University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）；Columbia University（哥伦比亚大学）；Massachusetts Institute of Technology（麻省理工学院）；Harvard University（哈佛大学）
University of Toronto（多伦多大学）；Vector Institute（向量研究所）；TU Wien（维也纳工业大学）；University of Pennsylvania（宾夕法尼亚大学）；University of Warsaw（华沙大学）；UC Berkeley（加州大学伯克利分校）；Polytechnique Montréal（蒙特利尔综合理工学院）；Mila – Quebec AI Institute（Mila – 魁北克人工智能研究所）
Relative Entropy Regularized Reinforcement Learning for Efficient Encrypted Policy Synthesis
School of Aeronautics and Astronautics, Purdue University（航空宇航工程学院，普渡大学）；ASRI, the Department of Electrical and Computer Engineering, Seoul National University（ASRI，电气电子工程系，首尔国立大学）；Japan Society for the Promotion of Science（日本学术振兴会）
