Reformulate LLM Reinforcement Learning for Efficient Training under Black-box Discrepancy

• URL: http://arxiv.org/abs/2606.08779v2
• Date: Tue, 09 Jun 2026 07:53:57 GMT
• Title: Reformulate LLM Reinforcement Learning for Efficient Training under Black-box Discrepancy
• Abstract summary: Reinforcement Learning frequently suffers from unpredictable sub-optimum performance or even training collapses.<n>We find that the training policy can actively self-correct such a discrepancy when provided with an appropriate learning signal.<n>We formulate this problem as a Discrepancy-Constrained Markov Decision Process (DCMDP), where reward is coupled with a constraint that aligns trainingInference behavior.
• Score: 27.237699512463475
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Reinforcement Learning (RL) has emerged as a pivotal post-training paradigm, yet it frequently suffers from unpredictable sub-optimum performance or even training collapses. Recent findings attribute these failures to a hidden train-inference discrepancy (or mismatch), stemming from the disparate underlying engines and architecture. We find that the training policy can actively self-correct such a discrepancy when provided with an appropriate learning signal. Then, we further empirically identify a discrepancy tolerance region: within this region, aggressively narrowing the discrepancy can suppress policy exploration and reduce learning efficiency, whereas outside this region, reducing excessive discrepancy improves optimization consistency and raises the achievable local performance ceiling. According to such findings, we formulate this problem as a Discrepancy-Constrained Markov Decision Process (DCMDP), where reward maximization is coupled with a constraint that aligns training-Inference behavior, achieving stable dual-objective optimization. To adaptively balance performance improvement and discrepancy control, we introduce a Lagrangian relaxation mechanism that dynamically adjusts the relative weight of the two objectives according to the current degree of discrepancy violation. This enables stable dual-objective optimization: the policy is allowed to explore freely within the tolerance region, while being guided back when the discrepancy exceeds the safe boundary. Empirically, DCMDP significantly improves the performance of 8B dense model (Qwen-3-8b) and 30B Mixture-of-Expert model (Qwen-3-30bA3b), and enables a heterogeneous training paradigm, where LLMs can be optimized in high-fidelity training setup while being explicitly aligned for low-cost, resource-constrained inference deployment.

Related papers

• ARMOR: Stabilizing On-Policy LLM RL with Off-Policy Anchor Samples [65.74946647574696]
  Reinforcement learning (RL) has significantly enhanced the reasoning capabilities of large language models (LLMs)<n>In this work, we investigate a critical source of this instability: over-optimization.<n>We propose ARMOR, a framework that shifts the paradigm from passive penalty to active sample stabilization.
  arXiv  Detail & Related papers  (2026-07-11T21:22:46Z)
• The Mirage of Optimizing Training Policies: Monotonic Inference Policies as the Real Objective for LLM Reinforcement Learning [65.72688364382331]
  Reinforcement learning (RL) has gained growing attention in large language model (LLM) post-training, yet RL training remains fragile and can suffer from instability or collapse.<n>One vital cause is training-inference mismatch: LLM adopts separate inference and training engines for generation efficiency and training precision.<n>We introduce Monotonic Inference Policy Update (MIPU), a two-step LLM RL framework that constructs sampler-referenced candidate updates and selectively accepts synchronized candidates.
  arXiv  Detail & Related papers  (2026-06-28T17:40:02Z)
• Uncertainty-Aware LLM-Guided Policy Shaping for Sparse-Reward Reinforcement Learning [2.384534878752428]
  We propose Uncertainty-Aware LLM-Guided Policy Shaping (ULPS), a framework that integrates a calibrated Large Language Model into the Reinforcement Learning training loop.<n>ULPS employs an A*-based oracle to synthesize optimal symbolic trajectories, which are used to fine-tune a BERT-based language model.<n>We observe consistent improvements in success rate, reward efficiency, and sample complexity over unguided, uncalibrated, and standard RL baselines.
  arXiv  Detail & Related papers  (2026-06-04T19:46:45Z)
• Post-Hoc Robustness for Model-Based Reinforcement Learning [1.4072254177584387]
  This work introduces post-hoc robustification of deep RL agents at inference time.<n>The goal is to improve robustness without any additional training of neural networks.
  arXiv  Detail & Related papers  (2026-06-02T11:43:13Z)
• Learning from Failures: Correction-Oriented Policy Optimization with Verifiable Rewards [73.44333771806282]
  Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as an effective paradigm for improving the reasoning capabilities of large language models.<n>We propose Correction-Oriented Policy Optimization (CIPO), a simple and effective extension to RLVR that converts on-policy failed trajectories into correction-oriented supervision.<n>CIPO improves learning effectiveness while explicitly enhancing the model's ability to correct its own errors.
  arXiv  Detail & Related papers  (2026-05-14T08:22:21Z)
• GIRL: Generative Imagination Reinforcement Learning via Information-Theoretic Hallucination Control [0.0]
  We introduce GIRL (Generative Imagination Reinforcement Learning), a latent world-model framework that addresses this failure mode with two key components.<n> GIRL reduces latent rollout drift by 38 to 61 percent across tasks relative to DreamerV3, improves return, and requires fewer environment interactions on long-horizon tasks.<n>A distilled-prior variant reduces overhead and improves computational efficiency relative to the full model.
  arXiv  Detail & Related papers  (2026-04-08T17:14:21Z)
• Rethinking the Trust Region in LLM Reinforcement Learning [72.25890308541334]
  Proximal Policy Optimization (PPO) serves as the de facto standard algorithm for Large Language Models (LLMs)<n>We propose Divergence Proximal Policy Optimization (DPPO), which substitutes clipping with a more principled constraint.<n>DPPO achieves superior training and efficiency compared to existing methods, offering a more robust foundation for RL-based fine-tuning.
  arXiv  Detail & Related papers  (2026-02-04T18:59:04Z)
• A Step Back: Prefix Importance Ratio Stabilizes Policy Optimization [58.116300485427764]
  Reinforcement learning post-training can elicit reasoning behaviors in large language models.<n> token-level correction often leads to unstable training dynamics when the degree of off-policyness is large.<n>We propose a simple yet effective objective, Minimum Prefix Ratio (MinPRO)
  arXiv  Detail & Related papers  (2026-01-30T08:47:19Z)
• CurES: From Gradient Analysis to Efficient Curriculum Learning for Reasoning LLMs [53.749193998004166]
  Curriculum learning plays a crucial role in enhancing the training efficiency of large language models.<n>We propose CurES, an efficient training method that accelerates convergence and employs Bayesian posterior estimation to minimize computational overhead.
  arXiv  Detail & Related papers  (2025-10-01T15:41:27Z)
• Stabilizing Policy Gradients for Sample-Efficient Reinforcement Learning in LLM Reasoning [77.92320830700797]
  Reinforcement Learning has played a central role in enabling reasoning capabilities of Large Language Models.<n>We propose a tractable computational framework that tracks and leverages curvature information during policy updates.<n>The algorithm, Curvature-Aware Policy Optimization (CAPO), identifies samples that contribute to unstable updates and masks them out.
  arXiv  Detail & Related papers  (2025-10-01T12:29:32Z)
• Distributionally Robust Federated Learning with Outlier Resilience [8.69285602685459]
  We study distributionally robust federated learning with explicit outlier resilience.<n>We reformulate the problem as a tractable Lagrangian penalty optimization, which admits robustness certificates.<n>Building on this reformulation, we propose the distributionally outlier-robust federated learning algorithm and establish its convergence guarantees.
  arXiv  Detail & Related papers  (2025-09-29T08:42:12Z)
• Binarization-Aware Adjuster: Bridging Continuous Optimization and Binary Inference in Edge Detection [0.0]
  Image edge detection (ED) faces a fundamental mismatch between training and inference.<n>In this paper, we propose a theoretical method to design a Binarization-Aware (BAA)<n>BAA explicitly incorporates binarization behavior into gradient-based optimization.
  arXiv  Detail & Related papers  (2025-06-14T11:56:44Z)
• Attribute Controlled Fine-tuning for Large Language Models: A Case Study on Detoxification [76.14641982122696]
  We propose a constraint learning schema for fine-tuning Large Language Models (LLMs) with attribute control. We show that our approach leads to an LLM that produces fewer inappropriate responses while achieving competitive performance on benchmarks and a toxicity detection task.
  arXiv  Detail & Related papers  (2024-10-07T23:38:58Z)

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.