English
Related papers

Related papers: Bootstrapped Mixed Rewards for RL Post-Training: I…

200 papers

Large language models excel at short-horizon reasoning tasks, but performance drops as reasoning horizon lengths increase. Existing approaches to combat this rely on inference-time scaffolding or costly step-level supervision, neither of…

Reinforcement learning (RL) is effective in enhancing the accuracy of large language models in complex reasoning tasks. Existing RL policy optimization frameworks rely on final-answer correctness as feedback signals and rarely capture the…

Artificial Intelligence · Computer Science 2026-04-13 Jinghan Zhang , Fengran Mo , Tharindu Cyril Weerasooriya , Ruimin Dai , Xiaoyan Han , Yanjie Fu , Dakuo Wang , Kunpeng Liu

We study reinforcement learning (RL) problems in which agents observe the reward or transition realizations at their current state before deciding which action to take. Such observations are available in many applications, including…

Machine Learning · Computer Science 2024-10-22 Nadav Merlis

Reinforcement learning (RL) has proven remarkably effective at improving the accuracy of language models in verifiable and deterministic domains like mathematics. Here, we examine if current RL methods are also effective at optimizing…

Machine Learning · Computer Science 2025-08-19 Michael Bereket , Jure Leskovec

Recent curriculum techniques in the post-training stage of LLMs have been empirically observed to outperform non-curriculum approaches in improving reasoning performance, yet a principled understanding of their effectiveness and limitations…

Machine Learning · Computer Science 2026-05-05 Dake Bu , Wei Huang , Andi Han , Atsushi Nitanda , Hau-San Wong , Qingfu Zhang , Taiji Suzuki

We study a theory of reinforcement learning (RL) in which the learner receives binary feedback only once at the end of an episode. While this is an extreme test case for theory, it is also arguably more representative of real-world…

Machine Learning · Computer Science 2022-08-23 Niladri S. Chatterji , Aldo Pacchiano , Peter L. Bartlett , Michael I. Jordan

Multi-objective reinforcement learning (MORL) is the generalization of standard reinforcement learning (RL) approaches to solve sequential decision making problems that consist of several, possibly conflicting, objectives. Generally, in…

Artificial Intelligence · Computer Science 2019-10-08 Xi Chen , Ali Ghadirzadeh , Mårten Björkman , Patric Jensfelt

Can reinforcement learning with hard, verifiable rewards teach a compact language model to reason about physics, or does it primarily learn to pattern-match toward correct answers? We study this question by training a 1.5B-parameter…

Artificial Intelligence · Computer Science 2026-03-05 Tarjei Paule Hage , Markus J. Buehler

Reinforcement learning with human feedback for aligning large language models (LLMs) trains a reward model typically using ranking loss with comparison pairs.However, the training procedure suffers from an inherent problem: the uncontrolled…

Computation and Language · Computer Science 2024-09-19 Hang Zhou , Chenglong Wang , Yimin Hu , Tong Xiao , Chunliang Zhang , Jingbo Zhu

Recent extensive research has demonstrated that the enhanced reasoning capabilities acquired by models through Reinforcement Learning with Verifiable Rewards (RLVR) are primarily concentrated within the rank-1 components. Predicated on this…

Machine Learning · Computer Science 2026-05-08 Hao Ye , Jisheng Dang , Junfeng Fang , Bimei Wang , Yizhou Zhang , Ning Lv , Wencan Zhang , Hong Peng , Bin Hu , Tat-Seng Chua

Reinforcement Learning (RL) has been shown to be effective in many scenarios. However, it typically requires the exploration of a sufficiently large number of state-action pairs, some of which may be unsafe. Consequently, its application to…

Systems and Control · Electrical Eng. & Systems 2022-06-24 Yousef Emam , Gennaro Notomista , Paul Glotfelter , Zsolt Kira , Magnus Egerstedt

Post-training for reasoning of large language models (LLMs) increasingly relies on verifiable rewards: deterministic checkers that provide 0-1 correctness signals. While reliable, such binary feedback is brittle--many tasks admit partially…

Computation and Language · Computer Science 2025-10-20 Leitian Tao , Ilia Kulikov , Swarnadeep Saha , Tianlu Wang , Jing Xu , Sharon Li , Jason E Weston , Ping Yu

Transformers trained via Reinforcement Learning (RL) with outcome-based supervision can spontaneously develop the ability to generate intermediate reasoning steps (Chain-of-Thought). Yet the mechanism by which sparse rewards drive policy…

Machine Learning · Computer Science 2026-02-03 Yuval Ran-Milo , Yotam Alexander , Shahar Mendel , Nadav Cohen

Real-time control for robotics is a popular research area in the reinforcement learning community. Through the use of techniques such as reward shaping, researchers have managed to train online agents across a multitude of domains. Despite…

Robotics · Computer Science 2023-04-21 Mihai Anca , Jonathan D. Thomas , Dabal Pedamonti , Matthew Studley , Mark Hansen

Language models must be adapted to understand and follow user instructions. Reinforcement learning is widely used to facilitate this -- typically using fixed criteria such as "helpfulness" and "harmfulness". In our work, we instead propose…

Computation and Language · Computer Science 2025-12-02 Vijay Viswanathan , Yanchao Sun , Shuang Ma , Xiang Kong , Meng Cao , Graham Neubig , Tongshuang Wu

Reinforcement learning (RL), particularly in sparse reward settings, often requires prohibitively large numbers of interactions with the environment, thereby limiting its applicability to complex problems. To address this, several prior…

Machine Learning · Computer Science 2020-11-20 Prasoon Goyal , Scott Niekum , Raymond J. Mooney

Although outcome-based reinforcement learning (RL) significantly advances the mathematical reasoning capabilities of Large Language Models (LLMs), its reliance on computationally expensive ground-truth annotations imposes a severe…

Machine Learning · Computer Science 2026-03-18 Zelin Zhang , Fei Cheng , Chenhui Chu

As machine learning has moved towards leveraging large models as priors for downstream tasks, the community has debated the right form of prior for solving reinforcement learning (RL) problems. If one were to try to prefetch as much…

Machine Learning · Computer Science 2026-02-13 Chongyi Zheng , Royina Karegoudra Jayanth , Benjamin Eysenbach

When applying reinforcement learning from human feedback (RLHF), the reward is learned from data and, therefore, always has some error. It is common to mitigate this by regularizing the policy with KL divergence from a base model, with the…

Machine Learning · Computer Science 2024-11-11 Thomas Kwa , Drake Thomas , Adrià Garriga-Alonso

Reinforcement learning has emerged as a powerful paradigm for unlocking reasoning capabilities in language models. However, relying on sparse rewards makes this process highly sample-inefficient, as models must navigate vast search spaces…

Machine Learning · Computer Science 2026-05-11 Ilia Mahrooghi , Aryo Lotfi , Emmanuel Abbe