Toward Evaluative Thinking: Meta Policy Optimization with Evolving Reward Models

• URL: http://arxiv.org/abs/2504.20157v1
• Date: Mon, 28 Apr 2025 18:02:35 GMT
• Title: Toward Evaluative Thinking: Meta Policy Optimization with Evolving Reward Models
• Authors: Zae Myung Kim, Chanwoo Park, Vipul Raheja, Dongyeop Kang,
• Abstract summary: We introduce a framework that integrates a meta-reward model that dynamically refines the reward model's prompt throughout training.<n>MPO promotes a more stable policy optimization, and greatly reduces the need for manual reward prompt design.<n>It yields performance on par with or better than models guided by extensively hand-crafted reward prompts.
• Score: 21.781693384336567
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Reward-based alignment methods for large language models (LLMs) face two key limitations: vulnerability to reward hacking, where models exploit flaws in the reward signal; and reliance on brittle, labor-intensive prompt engineering when LLMs are used as reward models. We introduce Meta Policy Optimization (MPO), a framework that addresses these challenges by integrating a meta-reward model that dynamically refines the reward model's prompt throughout training. In MPO, the meta-reward model monitors the evolving training context and continuously adjusts the reward model's prompt to maintain high alignment, providing an adaptive reward signal that resists exploitation by the policy. This meta-learning approach promotes a more stable policy optimization, and greatly reduces the need for manual reward prompt design. It yields performance on par with or better than models guided by extensively hand-crafted reward prompts. Furthermore, we show that MPO maintains its effectiveness across diverse tasks, such as question answering and mathematical reasoning, without requiring specialized reward designs. Beyond standard RLAIF, MPO's meta-learning formulation is readily extensible to higher-level alignment frameworks. Overall, this method addresses theoretical and practical challenges in reward-based RL alignment for LLMs, paving the way for more robust and adaptable alignment strategies. The code and models will be publicly shared.

Related papers

• Adversarial Training of Reward Models [74.17196154247964]
  We introduce Adv-RM, a novel adversarial training framework that automatically identifies adversarial examples.<n>By leveraging reinforcement learning, Adv-RM trains a policy to expose vulnerabilities in large state-of-the-art reward models.<n>We demonstrate that Adv-RM significantly outperforms conventional reward training.
  arXiv  Detail & Related papers  (2025-04-08T15:38:25Z)
• Approximated Variational Bayesian Inverse Reinforcement Learning for Large Language Model Alignment [0.618727087412292]
  The alignment of large language models (LLMs) is crucial for generating helpful and harmless content. Existing approaches leverage preference-based human feedback data to learn the reward function. We propose a novel training objective, Approximated Variational Alignment (AVA), to perform LLM alignment through Approximated Variational Reward Learning (AVRIL)
  arXiv  Detail & Related papers  (2024-11-14T10:37:34Z)
• HAF-RM: A Hybrid Alignment Framework for Reward Model Training [51.59246299566669]
  We propose a hybrid alignment framework HaF-RM for reward model training. It offers a principled and effective approach to enhancing the performance and alignment of reward models.
  arXiv  Detail & Related papers  (2024-07-04T23:26:56Z)
• Robust Model-Based Reinforcement Learning with an Adversarial Auxiliary Model [2.9109581496560044]
  An RL agent that trains in a certain Markov decision process (MDP) often struggles to perform well in nearly identical MDPs. We employ the framework of Robust MDPs in a model-based setting and introduce a novel learned transition model. Our experimental results indicate a notable improvement in policy robustness on high-dimensional MuJoCo control tasks.
  arXiv  Detail & Related papers  (2024-06-14T12:37:08Z)
• ALaRM: Align Language Models via Hierarchical Rewards Modeling [41.79125107279527]
  We introduce ALaRM, the first framework modeling hierarchical rewards in reinforcement learning from human feedback. The framework addresses the limitations of current alignment approaches, by integrating holistic rewards with aspect-specific rewards. We validate our approach through applications in long-form question answering and machine translation tasks.
  arXiv  Detail & Related papers  (2024-03-11T14:28:40Z)
• Entropy-Regularized Token-Level Policy Optimization for Language Agent Reinforcement [67.1393112206885]
  Large Language Models (LLMs) have shown promise as intelligent agents in interactive decision-making tasks. We introduce Entropy-Regularized Token-level Policy Optimization (ETPO), an entropy-augmented RL method tailored for optimizing LLMs at the token level. We assess the effectiveness of ETPO within a simulated environment that models data science code generation as a series of multi-step interactive tasks.
  arXiv  Detail & Related papers  (2024-02-09T07:45:26Z)
• Let's reward step by step: Step-Level reward model as the Navigators for Reasoning [64.27898739929734]
  Process-Supervised Reward Model (PRM) furnishes LLMs with step-by-step feedback during the training phase. We propose a greedy search algorithm that employs the step-level feedback from PRM to optimize the reasoning pathways explored by LLMs. To explore the versatility of our approach, we develop a novel method to automatically generate step-level reward dataset for coding tasks and observed similar improved performance in the code generation tasks.
  arXiv  Detail & Related papers  (2023-10-16T05:21:50Z)
• Reparameterized Policy Learning for Multimodal Trajectory Optimization [61.13228961771765]
  We investigate the challenge of parametrizing policies for reinforcement learning in high-dimensional continuous action spaces. We propose a principled framework that models the continuous RL policy as a generative model of optimal trajectories. We present a practical model-based RL method, which leverages the multimodal policy parameterization and learned world model.
  arXiv  Detail & Related papers  (2023-07-20T09:05:46Z)
• When to Update Your Model: Constrained Model-based Reinforcement Learning [50.74369835934703]
  We propose a novel and general theoretical scheme for a non-decreasing performance guarantee of model-based RL (MBRL) Our follow-up derived bounds reveal the relationship between model shifts and performance improvement. A further example demonstrates that learning models from a dynamically-varying number of explorations benefit the eventual returns.
  arXiv  Detail & Related papers  (2022-10-15T17:57:43Z)
• Model-based Meta Reinforcement Learning using Graph Structured Surrogate Models [40.08137765886609]
  We show that our model, called a graph structured surrogate model (GSSM), outperforms state-of-the-art methods in predicting environment dynamics. Our approach is able to obtain high returns, while allowing fast execution during deployment by avoiding test time policy gradient optimization.
  arXiv  Detail & Related papers  (2021-02-16T17:21:55Z)
• Hindsight Expectation Maximization for Goal-conditioned Reinforcement Learning [26.631740480100724]
  We propose a graphical model framework for goal-conditioned RL, with an EM algorithm that operates on the lower bound of the RL objective. The E-step provides a natural interpretation of how 'learning in hindsight' techniques, such as HER, to handle extremely sparse goal-conditioned rewards. The M-step reduces policy optimization to supervised learning updates, which greatly stabilizes end-to-end training on high-dimensional inputs such as images.
  arXiv  Detail & Related papers  (2020-06-13T03:25:31Z)

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.