Copertina

Reference-Sampled Boltzmann Projection for KL-Regularized RLVR: Target-Matched Weighted SFT, Finite One-Shot Gaps, and Policy Mirror Descent

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Shu, Yao, Wei, Chenxing, Lin, Hongbin, Qiu, Shuang, Xiong, Hui
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Machine Learning Artificial Intelligence
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

On the Effect of Regularization in Policy Mirror Descent
di: Kleuker, Jan Felix, et al.
Pubblicazione: (2025)

One-Step Flow Policy Mirror Descent
di: Chen, Tianyi, et al.
Pubblicazione: (2025)

Forward KL Regularized Preference Optimization for Aligning Diffusion Policies
di: Shan, Zhao, et al.
Pubblicazione: (2024)

Flow Matching Policy Optimization with Mirror Descent and Entropy Constraints
di: Gao, Ting, et al.
Pubblicazione: (2026)

Policy Mirror Descent with Lookahead
di: Protopapas, Kimon, et al.
Pubblicazione: (2024)

On the Convergence of Policy in Unregularized Policy Mirror Descent
di: Lin, Dachao, et al.
Pubblicazione: (2022)

Sample Complexity of Neural Policy Mirror Descent for Policy Optimization on Low-Dimensional Manifolds
di: Xu, Zhenghao, et al.
Pubblicazione: (2023)

Functional Acceleration for Policy Mirror Descent
di: Chelu, Veronica, et al.
Pubblicazione: (2024)

Listwise Policy Optimization: Group-based RLVR as Target-Projection on the LLM Response Simplex
di: Qu, Yun, et al.
Pubblicazione: (2026)

Stochastic MeanFlow Policies: One-Step Generative Control with Entropic Mirror Descent
di: Wang, Zeyuan, et al.
Pubblicazione: (2026)

A New Kernel Regularity Condition for Distributed Mirror Descent: Broader Coverage and Simpler Analysis
di: Qiu, Junwen, et al.
Pubblicazione: (2026)

Decouple before Integration: Test-time Synthesis of SFT and RLVR Task Vectors
di: Yuan, Chaohao, et al.
Pubblicazione: (2026)

KL-Regularized RLHF with Multiple Reference Models: Exact Solutions and Sample Complexity
di: Aminian, Gholamali, et al.
Pubblicazione: (2025)

Entropy-Guided Multiplicative Updates: KL Projections for Multi-Factor Target Exposures
di: Qiu, Yimeng
Pubblicazione: (2025)

Stress-Aware Learning under KL Drift via Trust-Decayed Mirror Descent
di: Raj, Gabriel Nixon
Pubblicazione: (2025)

Approximation of Log-Partition Function in Policy Mirror Descent Induces Implicit Regularization for LLM Post-Training
di: Xu, Zhenghao, et al.
Pubblicazione: (2026)

Minimizing Weighted Counterfactual Regret with Optimistic Online Mirror Descent
di: Xu, Hang, et al.
Pubblicazione: (2024)

Policy Mirror Descent with Temporal Difference Learning: Sample Complexity under Online Markov Data
di: Li, Wenye, et al.
Pubblicazione: (2025)

On the Convergence of Policy Mirror Descent with Temporal Difference Evaluation
di: Liu, Jiacai, et al.
Pubblicazione: (2025)

Stability and Robustness via Regularization: Bandit Inference via Regularized Stochastic Mirror Descent
di: Halder, Budhaditya, et al.
Pubblicazione: (2026)

Fast Rates in $α$-Potential Games via Regularized Mirror Descent
di: Chen, Claire, et al.
Pubblicazione: (2026)

A Unified Approach to Controlling Implicit Regularization via Mirror Descent
di: Sun, Haoyuan, et al.
Pubblicazione: (2023)

Score-Regularized Joint Sampling with Importance Weights for Flow Matching
di: Liu, Xinshuang, et al.
Pubblicazione: (2025)

Convergence of Policy Mirror Descent Beyond Compatible Function Approximation
di: Sherman, Uri, et al.
Pubblicazione: (2025)

StaQ it! Growing neural networks for Policy Mirror Descent
di: Shilova, Alena, et al.
Pubblicazione: (2025)

Beyond State-Wise Mirror Descent: Offline Policy Optimization with Parametric Policies
di: Li, Xiang, et al.
Pubblicazione: (2026)

Energy-Weighted Flow Matching: Unlocking Continuous Normalizing Flows for Efficient and Scalable Boltzmann Sampling
di: Dern, Niclas, et al.
Pubblicazione: (2025)

Mirror Descent Methods with Weighting Scheme for Outputs for Constrained Variational Inequality Problems
di: Alkousa, Mohammad S., et al.
Pubblicazione: (2025)

Mirror Descent Policy Optimisation for Robust Constrained Markov Decision Processes
di: Bossens, David M., et al.
Pubblicazione: (2025)

Shorter but not Worse: Frugal Reasoning via Easy Samples as Length Regularizers in Math RLVR
di: Bounhar, Abdelaziz, et al.
Pubblicazione: (2025)

Taming Audio VAEs via Target-KL Regularization
di: Seetharaman, Prem, et al.
Pubblicazione: (2026)

Target Mirror Descent: A Unifying Framework for Solving Monotone Variational Inequalities
di: Chen, Yu-Wen, et al.
Pubblicazione: (2026)

Parameter-free Mirror Descent
di: Jacobsen, Andrew, et al.
Pubblicazione: (2022)

Mirror Descent on Riemannian Manifolds
di: Jiang, Jiaxin, et al.
Pubblicazione: (2026)

A Mirror Descent Perspective of Smoothed Sign Descent
di: Wang, Shuyang, et al.
Pubblicazione: (2024)

Shuffling the Stochastic Mirror Descent via Dual Lipschitz Continuity and Kernel Conditioning
di: Qiu, Junwen, et al.
Pubblicazione: (2026)

Finite-Particle Rates for Regularized Stein Variational Gradient Descent
di: He, Ye, et al.
Pubblicazione: (2026)

HEALing Entropy Collapse: Enhancing Exploration in Few-Shot RLVR via Hybrid-Domain Entropy Dynamics Alignment
di: Liu, Zhanyu, et al.
Pubblicazione: (2026)

A Novel Framework for Policy Mirror Descent with General Parameterization and Linear Convergence
di: Alfano, Carlo, et al.
Pubblicazione: (2023)

Leave No One Undermined: Policy Targeting with Regret Aversion
di: Kitagawa, Toru, et al.
Pubblicazione: (2025)