Image de couverture de livre

Controllable Exploration in Hybrid-Policy RLVR for Multi-Modal Reasoning

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Huang, Zhuoxu, Jia, Mengxi, Sun, Hao, Li, Xuelong, Han, Jungong
Format:	Preprint
Publié:	2026
Sujets:	Machine Learning Artificial Intelligence
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Segment-Aligned Policy Optimization for Multi-Modal Reasoning
par: Gao, Lei, et autres
Publié: (2026)

Depth-Breadth Synergy in RLVR: Unlocking LLM Reasoning Gains with Adaptive Exploration
par: Yang, Zhicheng, et autres
Publié: (2025)

Generalization of RLVR Using Causal Reasoning as a Testbed
par: Lu, Brian, et autres
Publié: (2025)

HEALing Entropy Collapse: Enhancing Exploration in Few-Shot RLVR via Hybrid-Domain Entropy Dynamics Alignment
par: Liu, Zhanyu, et autres
Publié: (2026)

Quantile Advantage Estimation: Stabilizing RLVR for LLM Reasoning
par: Wu, Junkang, et autres
Publié: (2025)

RLVR without Ineffective Samples: Group Prioritized Off-Policy Optimization for LLM Reasoning
par: Mao, Yixiu, et autres
Publié: (2026)

On the Direction of RLVR Updates for LLM Reasoning: Identification and Exploitation
par: Huang, Kexin, et autres
Publié: (2026)

Parameter-Efficient Fine-Tuning for Continual Learning: A Neural Tangent Kernel Perspective
par: Liu, Jingren, et autres
Publié: (2024)

SAGE: Shaping Anchors for Guided Exploration in RLVR of LLMs
par: Lee, Chanuk, et autres
Publié: (2026)

Multi-Modal Manipulation via Multi-Modal Policy Consensus
par: Chen, Haonan, et autres
Publié: (2025)

Exploration vs Exploitation: Rethinking RLVR through Clipping, Entropy, and Spurious Reward
par: Chen, Peter, et autres
Publié: (2025)

ORBIT: On-policy Exploration-Exploitation for Controllable Multi-Budget Reasoning
par: Liang, Kun, et autres
Publié: (2026)

Understanding and Preventing Entropy Collapse in RLVR with On-Policy Entropy Flow Optimization
par: Xu, Huimin, et autres
Publié: (2026)

SIME: Enhancing Policy Self-Improvement with Modal-level Exploration
par: Jin, Yang, et autres
Publié: (2025)

Monitorability as a Free Gift: How RLVR Spontaneously Aligns Reasoning
par: Xiong, Zidi, et autres
Publié: (2026)

JudgeRLVR: Judge First, Generate Second for Efficient Reasoning
par: Duo, Jiangshan, et autres
Publié: (2026)

Listwise Policy Optimization: Group-based RLVR as Target-Projection on the LLM Response Simplex
par: Qu, Yun, et autres
Publié: (2026)

CLIPO: Contrastive Learning in Policy Optimization Generalizes RLVR
par: Cui, Sijia, et autres
Publié: (2026)

GeoRA: Geometry-Aware Low-Rank Adaptation for RLVR
par: Zhang, Jiaying, et autres
Publié: (2026)

Adaptive Negative Reinforcement for LLM Reasoning:Dynamically Balancing Correction and Diversity in RLVR
par: Ingle, Yash, et autres
Publié: (2026)

The Path Not Taken: RLVR Provably Learns Off the Principals
par: Zhu, Hanqing, et autres
Publié: (2025)

Categorical Policies: Multimodal Policy Learning and Exploration in Continuous Control
par: Islam, SM Mazharul, et autres
Publié: (2025)

Adaptive Exploration for Multi-Reward Multi-Policy Evaluation
par: Russo, Alessio, et autres
Publié: (2025)

Shorter but not Worse: Frugal Reasoning via Easy Samples as Length Regularizers in Math RLVR
par: Bounhar, Abdelaziz, et autres
Publié: (2025)

Rethinking Entropy Interventions in RLVR: An Entropy Change Perspective
par: Hao, Zhezheng, et autres
Publié: (2025)

IRDS: Interpretable RLVR Data Selection via Verifier-Coupled Sparse Autoencoder Coverage
par: Li, Yuhan, et autres
Publié: (2026)

On the Implicit Reward Overfitting and the Low-rank Dynamics in RLVR
par: Ye, Hao, et autres
Publié: (2026)

Conformal Selective Acting: Anytime-Valid Risk Control for RLVR-Trained LLMs
par: Khosravi, Hamed, et autres
Publié: (2026)

Part II: ROLL Flash -- Accelerating RLVR and Agentic Training with Asynchrony
par: Lu, Han, et autres
Publié: (2025)

Spurious Rewards: Rethinking Training Signals in RLVR
par: Shao, Rulin, et autres
Publié: (2025)

Towards Robust Multi-Modal Reasoning via Model Selection
par: Liu, Xiangyan, et autres
Publié: (2023)

A State-of-the-Art SQL Reasoning Model using RLVR
par: Ali, Alnur, et autres
Publié: (2025)

The Debate on RLVR Reasoning Capability Boundary: Shrinkage, Expansion, or Both? A Two-Stage Dynamic View
par: Yao, Xinhao, et autres
Publié: (2025)

DAMBench: A Multi-Modal Benchmark for Deep Learning-based Atmospheric Data Assimilation
par: Wang, Hao, et autres
Publié: (2025)

Quantifying Empirical Compute-Supervision Tradeoffs in RLVR
par: Mitsuhashi, Ryo, et autres
Publié: (2026)

VL Norm: Rethink Loss Aggregation in RLVR
par: He, Zhiyuan, et autres
Publié: (2025)

Flexible Entropy Control in RLVR with a Gradient-Preserving Perspective
par: Chen, Kun, et autres
Publié: (2026)

Asymmetric Advantage Modulation Calibrates Entropy Dynamics in RLVR
par: Gu, Hengrui, et autres
Publié: (2026)

Towards Learnable Anchor for Deep Multi-View Clustering
par: Wang, Bocheng, et autres
Publié: (2025)

Exploiting Hybrid Policy in Reinforcement Learning for Interpretable Temporal Logic Manipulation
par: Zhang, Hao, et autres
Publié: (2024)