Improving DAPO from a Mixed-Policy Perspective

• URL: http://arxiv.org/abs/2507.12931v3
• Date: Tue, 19 Aug 2025 06:04:15 GMT
• Title: Improving DAPO from a Mixed-Policy Perspective
• Authors: Hongze Tan, Yuchen Li,
• Abstract summary: This paper introduces two novel modifications to the Dynamic sAmpling Policy Optimization (DAPO) algorithm.<n>We first propose a method that incorporates a pre-trained, stable guiding policy to provide off-policy experience.<n>We then extend this idea to re-utilize zero-reward samples, which are often discarded by dynamic sampling strategies.
• Score: 6.237966553429477
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: This paper introduces two novel modifications to the Dynamic sAmpling Policy Optimization (DAPO) algorithm [1], approached from a mixed-policy perspective. Standard policy gradient methods can suffer from instability and sample inefficiency, particularly in sparse reward settings. To address this, we first propose a method that incorporates a pre-trained, stable guiding policy ($\piphi$) to provide off-policy experience, thereby regularizing the training of the target policy ($\pion$). This approach improves training stability and convergence speed by adaptively adjusting the learning step size. Secondly, we extend this idea to re-utilize zero-reward samples, which are often discarded by dynamic sampling strategies like DAPO's. By treating these samples as a distinct batch guided by the expert policy, we further enhance sample efficiency. We provide a theoretical analysis for both methods, demonstrating that their objective functions converge to the optimal solution within the established theoretical framework of reinforcement learning. The proposed mixed-policy framework effectively balances exploration and exploitation, promising more stable and efficient policy optimization.

Related papers

• ExO-PPO: an Extended Off-policy Proximal Policy Optimization Algorithm [2.6813717321945103]
  We propose a new PPO variant based on the stability guarantee from conservative on-policy iteration with a more efficient off-policy data utilization.<n>Compared with PPO and some other state-of-the-art variants, we demonstrate an improved performance of ExO-PPO with balanced sample efficiency and stability on varied tasks.
  arXiv  Detail & Related papers  (2026-02-10T12:29:57Z)
• Behaviour Policy Optimization: Provably Lower Variance Return Estimates for Off-Policy Reinforcement Learning [52.97053840476386]
  We show that well-designed behaviour policies can be used to collect off-policy data for provably lower variance return estimates.<n>We extend this key insight to the online reinforcement learning setting, where both policy evaluation and improvement are interleaved.
  arXiv  Detail & Related papers  (2025-11-13T23:06:40Z)
• ACPO: Adaptive Curriculum Policy Optimization for Aligning Vision-Language Models in Complex Reasoning [17.928214942495412]
  ACPO employs a dynamic curriculum that orchestrates a principled transition from a stable, near on-policy exploration phase to an efficient, off-policy exploitation phase.<n>We conduct extensive experiments on a suite of challenging multimodal reasoning benchmarks, including MathVista, LogicVista, and MMMU-Pro.<n>Results demonstrate that ACPO consistently outperforms strong baselines such as DAPO and PAPO, achieving state-of-the-art performance, accelerated convergence, and superior training stability.
  arXiv  Detail & Related papers  (2025-10-01T09:11:27Z)
• Polychromic Objectives for Reinforcement Learning [63.37185057794815]
  Reinforcement learning fine-tuning (RLFT) is a dominant paradigm for improving pretrained policies for downstream tasks.<n>We introduce an objective for policy methods that explicitly enforces the exploration and refinement of diverse generations.<n>We show how proximal policy optimization (PPO) can be adapted to optimize this objective.
  arXiv  Detail & Related papers  (2025-09-29T19:32:11Z)
• Achieving $\widetilde{\mathcal{O}}(\sqrt{T})$ Regret in Average-Reward POMDPs with Known Observation Models [56.92178753201331]
  We tackle average-reward infinite-horizon POMDPs with an unknown transition model.<n>We present a novel and simple estimator that overcomes this barrier.
  arXiv  Detail & Related papers  (2025-01-30T22:29:41Z)
• Optimal Policy Adaptation under Covariate Shift [15.703626346971182]
  We propose principled approaches for learning the optimal policy in the target domain by leveraging two datasets.<n>We derive the identifiability assumptions for the reward induced by a given policy.<n>We then learn the optimal policy by optimizing the estimated reward.
  arXiv  Detail & Related papers  (2025-01-14T12:33:02Z)
• Reflective Policy Optimization [20.228281670899204]
  Reflective Policy Optimization (RPO) amalgamates past and future state-action information for policy optimization. RPO empowers the agent for introspection, allowing modifications to its actions within the current state. Empirical results demonstrate RPO's feasibility and efficacy in two reinforcement learning benchmarks.
  arXiv  Detail & Related papers  (2024-06-06T01:46:49Z)
• Clipped-Objective Policy Gradients for Pessimistic Policy Optimization [3.2996723916635275]
  Policy gradient methods seek to produce monotonic improvement through bounded changes in policy outputs. In this work, we find that the performance of PPO, when applied to continuous action spaces, may be consistently improved through a simple change in objective. We show that the clipped-objective policy gradient (COPG) objective is on average "pessimistic" compared to both the PPO objective and (2) this pessimism promotes enhanced exploration.
  arXiv  Detail & Related papers  (2023-11-10T03:02:49Z)
• Statistically Efficient Variance Reduction with Double Policy Estimation for Off-Policy Evaluation in Sequence-Modeled Reinforcement Learning [53.97273491846883]
  We propose DPE: an RL algorithm that blends offline sequence modeling and offline reinforcement learning with Double Policy Estimation. We validate our method in multiple tasks of OpenAI Gym with D4RL benchmarks.
  arXiv  Detail & Related papers  (2023-08-28T20:46:07Z)
• Acceleration in Policy Optimization [50.323182853069184]
  We work towards a unifying paradigm for accelerating policy optimization methods in reinforcement learning (RL) by integrating foresight in the policy improvement step via optimistic and adaptive updates. We define optimism as predictive modelling of the future behavior of a policy, and adaptivity as taking immediate and anticipatory corrective actions to mitigate errors from overshooting predictions or delayed responses to change. We design an optimistic policy gradient algorithm, adaptive via meta-gradient learning, and empirically highlight several design choices pertaining to acceleration, in an illustrative task.
  arXiv  Detail & Related papers  (2023-06-18T15:50:57Z)
• A Regularized Implicit Policy for Offline Reinforcement Learning [54.7427227775581]
  offline reinforcement learning enables learning from a fixed dataset, without further interactions with the environment. We propose a framework that supports learning a flexible yet well-regularized fully-implicit policy. Experiments and ablation study on the D4RL dataset validate our framework and the effectiveness of our algorithmic designs.
  arXiv  Detail & Related papers  (2022-02-19T20:22:04Z)
• Generalized Proximal Policy Optimization with Sample Reuse [8.325359814939517]
  We combine the theoretically supported stability benefits of on-policy algorithms with the sample efficiency of off-policy algorithms. We develop policy improvement guarantees that are suitable for the off-policy setting, and connect these bounds to the clipping mechanism used in Proximal Policy Optimization. This motivates an off-policy version of the popular algorithm that we call Generalized Proximal Policy Optimization with Sample Reuse.
  arXiv  Detail & Related papers  (2021-10-29T20:22:31Z)
• Off-policy Reinforcement Learning with Optimistic Exploration and Distribution Correction [73.77593805292194]
  We train a separate exploration policy to maximize an approximate upper confidence bound of the critics in an off-policy actor-critic framework. To mitigate the off-policy-ness, we adapt the recently introduced DICE framework to learn a distribution correction ratio for off-policy actor-critic training.
  arXiv  Detail & Related papers  (2021-10-22T22:07:51Z)
• Variance Reduction based Experience Replay for Policy Optimization [3.0790370651488983]
  Variance Reduction Experience Replay (VRER) is a framework for the selective reuse of relevant samples to improve policy gradient estimation. VRER forms the foundation of our sample efficient off-policy learning algorithm known as Policy Gradient with VRER.
  arXiv  Detail & Related papers  (2021-10-17T19:28:45Z)
• Stable Policy Optimization via Off-Policy Divergence Regularization [50.98542111236381]
  Trust Region Policy Optimization (TRPO) and Proximal Policy Optimization (PPO) are among the most successful policy gradient approaches in deep reinforcement learning (RL) We propose a new algorithm which stabilizes the policy improvement through a proximity term that constrains the discounted state-action visitation distribution induced by consecutive policies to be close to one another. Our proposed method can have a beneficial effect on stability and improve final performance in benchmark high-dimensional control tasks.
  arXiv  Detail & Related papers  (2020-03-09T13:05:47Z)

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.