How to Allocate, How to Learn? Dynamic Rollout Allocation and Advantage Modulation for Policy Optimization

• URL: http://arxiv.org/abs/2602.19208v1
• Date: Sun, 22 Feb 2026 14:38:24 GMT
• Title: How to Allocate, How to Learn? Dynamic Rollout Allocation and Advantage Modulation for Policy Optimization
• Authors: Yangyi Fang, Jiaye Lin, Xiaoliang Fu, Cong Qin, Haolin Shi, Chaowen Hu, Lu Pan, Ke Zeng, Xunliang Cai,
• Abstract summary: DynaMO is a theoretically-grounded dual-pronged optimization framework.<n>We develop gradient-aware advantage modulation grounded in theoretical analysis of gradient magnitude bounds.<n>Our framework compensates for gradient attenuation of high-confidence correct actions while utilizing entropy changes as computable indicators to stabilize excessive update magnitudes.
• Score: 14.087451720550597
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has proven effective for Large Language Model (LLM) reasoning, yet current methods face key challenges in resource allocation and policy optimization dynamics: (i) uniform rollout allocation ignores gradient variance heterogeneity across problems, and (ii) the softmax policy structure causes gradient attenuation for high-confidence correct actions, while excessive gradient updates may destabilize training. Therefore, we propose DynaMO, a theoretically-grounded dual-pronged optimization framework. At the sequence level, we prove that uniform allocation is suboptimal and derive variance-minimizing allocation from the first principle, establishing Bernoulli variance as a computable proxy for gradient informativeness. At the token level, we develop gradient-aware advantage modulation grounded in theoretical analysis of gradient magnitude bounds. Our framework compensates for gradient attenuation of high-confidence correct actions while utilizing entropy changes as computable indicators to stabilize excessive update magnitudes. Extensive experiments conducted on a diverse range of mathematical reasoning benchmarks demonstrate consistent improvements over strong RLVR baselines. Our implementation is available at: \href{https://anonymous.4open.science/r/dynamo-680E/README.md}{https://anonymous.4open.science/r/dynamo}.

Related papers

• Stabilizing Policy Optimization via Logits Convexity [59.242732612484474]
  We show that the convexity of the supervised fine-tuning loss with respect to model logits plays a key role in enabling stable training.<n>Motivated by this observation, we propose Logits Convex Optimization (LCO), a simple yet effective policy optimization framework.
  arXiv  Detail & Related papers  (2026-03-01T07:40:12Z)
• GradAlign: Gradient-Aligned Data Selection for LLM Reinforcement Learning [55.03441672267886]
  We propose GradAlign, a gradient-aligned data selection method for reinforcement learning.<n>We evaluate GradAlign across three data regimes: unreliable reward signals, distribution imbalance, and low-utility training corpus.
  arXiv  Detail & Related papers  (2026-02-25T01:54:50Z)
• 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)
• Alternating Reinforcement Learning for Rubric-Based Reward Modeling in Non-Verifiable LLM Post-Training [29.56905427210088]
  gradient-ARM is a framework that jointly optimize a rubric generator and a judge using reinforcement learning from preference feedback.<n>We show that gradient-ARM achieves state-of-the-art performance among baselines on benchmarks and significantly improves downstream policy alignment in both offline and online reinforcement learning settings.
  arXiv  Detail & Related papers  (2026-02-02T00:50:53Z)
• Not All Preferences Are Created Equal: Stability-Aware and Gradient-Efficient Alignment for Reasoning Models [52.48582333951919]
  We propose a dynamic framework designed to enhance alignment reliability by maximizing the Signal-to-Noise Ratio of policy updates.<n>SAGE (Stability-Aware Gradient Efficiency) integrates a coarse-grained curriculum mechanism that refreshes candidate pools based on model competence.<n> Experiments on multiple mathematical reasoning benchmarks demonstrate that SAGE significantly accelerates convergence and outperforms static baselines.
  arXiv  Detail & Related papers  (2026-02-01T12:56:10Z)
• 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)
• Dynamic Rank Reinforcement Learning for Adaptive Low-Rank Multi-Head Self Attention in Large Language Models [0.0]
  We propose Dynamic Rank Reinforcement Learning (DR-RL), a novel framework that adaptively optimize the low-rank factorization of Multi-Head Self-Attention (MHSA) in Large Language Models (LLMs)<n>DR-RL maintains downstream accuracy statistically equivalent to full-rank attention while significantly reducing Floating Point Operations (FLOPs)<n>This work bridges the gap between adaptive efficiency and theoretical rigor in MHSA, offering a principled, mathematically grounded alternative to rank reduction techniques in resource-constrained deep learning.
  arXiv  Detail & Related papers  (2025-12-17T21:09:19Z)
• OBLR-PO: A Theoretical Framework for Stable Reinforcement Learning [12.77713716713937]
  We provide a unified theoretical framework that characterizes the statistical properties of commonly used policy-gradient estimators.<n>We derive an adaptive learning-rate schedule governed by the signal-to-noise ratio (SNR) of gradients.<n>We further show that the variance-optimal baseline is a gradient-weighted estimator, offering a new principle for variance reduction.
  arXiv  Detail & Related papers  (2025-11-28T16:09:28Z)
• 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)
• Optimizing Chain-of-Thought Reasoners via Gradient Variance Minimization in Rejection Sampling and RL [20.177871969184004]
  Chain-of-thought (CoT) reasoning can be formalized as a latent variable problem, where the model needs to generate intermediate reasoning steps.<n>Prior approaches such as iterative reward-ranked fine-tuning fail to account for variability in difficulty and convergence behavior.<n>We propose GVMRAFT, a prompt-specific Dynamic Sample Allocation Strategy to minimize gradient variance under a computational budget constraint.
  arXiv  Detail & Related papers  (2025-05-05T06:26:00Z)
• Rao-Blackwell Gradient Estimators for Equivariant Denoising Diffusion [55.95767828747407]
  In domains such as molecular and protein generation, physical systems exhibit inherent symmetries that are critical to model.<n>We present a framework that reduces training variance and provides a provably lower-variance gradient estimator.<n>We also present a practical implementation of this estimator incorporating the loss and sampling procedure through a method we call Orbit Diffusion.
  arXiv  Detail & Related papers  (2025-02-14T03:26:57Z)
• Large Continual Instruction Assistant [59.585544987096974]
  Continual Instruction Tuning (CIT) is adopted to instruct Large Models to follow human intent data by data.<n>Existing update gradient would heavily destroy the performance on previous datasets during CIT process.<n>We propose a general continual instruction tuning framework to address the challenge.
  arXiv  Detail & Related papers  (2024-10-08T11:24:59Z)

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.