Imagen de Portada

The Rank and Gradient Lost in Non-stationarity: Sample Weight Decay for Mitigating Plasticity Loss in Reinforcement Learning

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Wu, Zihao, Tang, Hongyao, Ma, Yi, Liu, Jiashun, Zheng, Yan, Hao, Jianye
Formato:	Preprint
Publicado:	2026
Materias:	Machine Learning
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Mitigating Plasticity Loss in Continual Reinforcement Learning by Reducing Churn
por: Tang, Hongyao, et al.
Publicado: (2025)

Squeeze the Soaked Sponge: Efficient Off-policy Reinforcement Finetuning for Large Language Model
por: Liang, Jing, et al.
Publicado: (2025)

Can We Optimize Deep RL Policy Weights as Trajectory Modeling?
por: Tang, Hongyao
Publicado: (2025)

ENOTO: Improving Offline-to-Online Reinforcement Learning with Q-Ensembles
por: Zhao, Kai, et al.
Publicado: (2023)

MUVLA: Learning to Explore Object Navigation via Map Understanding
por: Han, Peilong, et al.
Publicado: (2025)

Prioritized Trajectory Replay: A Replay Memory for Data-driven Reinforcement Learning
por: Liu, Jinyi, et al.
Publicado: (2023)

Improving Deep Reinforcement Learning by Reducing the Chain Effect of Value and Policy Churn
por: Tang, Hongyao, et al.
Publicado: (2024)

Scaling DRL for Decision Making: A Survey on Data, Network, and Training Budget Strategies
por: Ma, Yi, et al.
Publicado: (2025)

SPHERE: Mitigating the Loss of Spectral Plasticity in Mixture-of-Experts for Deep Reinforcement Learning
por: Luo, Lirui, et al.
Publicado: (2026)

Embodied-R1: Reinforced Embodied Reasoning for General Robotic Manipulation
por: Yuan, Yifu, et al.
Publicado: (2025)

MFE-ETP: A Comprehensive Evaluation Benchmark for Multi-modal Foundation Models on Embodied Task Planning
por: Zhang, Min, et al.
Publicado: (2024)

On the Overlooked Pitfalls of Weight Decay and How to Mitigate Them: A Gradient-Norm Perspective
por: Xie, Zeke, et al.
Publicado: (2020)

Uni-RLHF: Universal Platform and Benchmark Suite for Reinforcement Learning with Diverse Human Feedback
por: Yuan, Yifu, et al.
Publicado: (2024)

ForceFlow: Learning to Feel and Act via Contact-Driven Flow Matching
por: Zhang, Shuoheng, et al.
Publicado: (2026)

Succeed or Learn Slowly: Sample Efficient Off-Policy Reinforcement Learning for Mobile App Control
por: Papoudakis, Georgios, et al.
Publicado: (2025)

Gradient conformal stationarity and the CMC condition in LRS spacetimes
por: Amery, Gareth, et al.
Publicado: (2024)

Adaptive Stochastic Gradient Descents on Manifolds with an Application on Weighted Low-Rank Approximation
por: Yang, Peiqi, et al.
Publicado: (2025)

Plasticity Loss in Deep Reinforcement Learning: A Survey
por: Klein, Timo, et al.
Publicado: (2024)

MODULI: Unlocking Preference Generalization via Diffusion Models for Offline Multi-Objective Reinforcement Learning
por: Yuan, Yifu, et al.
Publicado: (2024)

Advancing Autonomous VLM Agents via Variational Subgoal-Conditioned Reinforcement Learning
por: Wu, Qingyuan, et al.
Publicado: (2025)

The Lifecycle of the Spectral Edge: From Gradient Learning to Weight-Decay Compression
por: Xu, Yongzhong
Publicado: (2026)

XP-MARL: Auxiliary Prioritization in Multi-Agent Reinforcement Learning to Address Non-Stationarity
por: Xu, Jianye, et al.
Publicado: (2024)

SigmaRL: A Sample-Efficient and Generalizable Multi-Agent Reinforcement Learning Framework for Motion Planning
por: Xu, Jianye, et al.
Publicado: (2024)

Weighted Temporal Decay Loss for Learning Wearable PPG Data with Sparse Clinical Labels
por: Chung, Yunsung, et al.
Publicado: (2026)

A Study of Plasticity Loss in On-Policy Deep Reinforcement Learning
por: Juliani, Arthur, et al.
Publicado: (2024)

Ranking-aware Reinforcement Learning for Ordinal Ranking
por: Hao, Aiming, et al.
Publicado: (2026)

Neuroplastic Expansion in Deep Reinforcement Learning
por: Liu, Jiashun, et al.
Publicado: (2024)

Comparative analysis of stationarity for Bitcoin and the S&P500
por: Tang, Yaoyue, et al.
Publicado: (2024)

DistRL: An Asynchronous Distributed Reinforcement Learning Framework for On-Device Control Agents
por: Wang, Taiyi, et al.
Publicado: (2024)

Coefficients-Preserving Sampling for Reinforcement Learning with Flow Matching
por: Wang, Feng, et al.
Publicado: (2025)

Mitigating Voltage Decay via Chemo‐Mechanical Decoupling Enabled by a Rigid‐Soft Gradient Interphase in Li‐Rich Layered Oxides
por: Lingcai Zeng, et al.
Publicado: (2026)

Loss- and Reward-Weighting for Efficient Distributed Reinforcement Learning
por: Holen, Martin, et al.
Publicado: (2023)

SGD and Weight Decay Secretly Minimize the Rank of Your Neural Network
por: Galanti, Tomer, et al.
Publicado: (2022)

Lost in Embeddings: Information Loss in Vision-Language Models
por: Li, Wenyan, et al.
Publicado: (2025)

Quasi-stationarity of the Dyson Brownian motion with collisions
por: Guillin, Arnaud, et al.
Publicado: (2025)

MENTOR: Guiding Hierarchical Reinforcement Learning with Human Feedback and Dynamic Distance Constraint
por: Zhou, Xinglin, et al.
Publicado: (2024)

Curl Descent: Non-Gradient Learning Dynamics with Sign-Diverse Plasticity
por: Ninou, Hugo, et al.
Publicado: (2025)

Revisiting Clustering of Neural Bandits: Selective Reinitialization for Mitigating Loss of Plasticity
por: Su, Zhiyuan, et al.
Publicado: (2025)

Lost in the Passage: Passage-level In-context Learning Does Not Necessarily Need a "Passage"
por: Sun, Hao, et al.
Publicado: (2025)

Non-stationarity Characteristics in Dynamic Vehicular ISAC Channels at 28 GHz
por: Zhang, Zhengyu, et al.
Publicado: (2024)