Simplifying deep temporal difference learning
Q-learning played a foundational role in the field reinforcement learning (RL). However, TD
algorithms with off-policy data, such as Q-learning, or nonlinear function approximation like …
algorithms with off-policy data, such as Q-learning, or nonlinear function approximation like …
Model Surgery: Modulating LLM's Behavior Via Simple Parameter Editing
Large Language Models (LLMs) have demonstrated great potential as generalist assistants,
showcasing powerful task understanding and problem-solving capabilities. To deploy LLMs …
showcasing powerful task understanding and problem-solving capabilities. To deploy LLMs …
Reinforcement Learning Gradients as Vitamin for Online Finetuning Decision Transformers
K Yan, AG Schwing, YX Wang - arXiv preprint arXiv:2410.24108, 2024 - arxiv.org
Decision Transformers have recently emerged as a new and compelling paradigm for offline
Reinforcement Learning (RL), completing a trajectory in an autoregressive way. While …
Reinforcement Learning (RL), completing a trajectory in an autoregressive way. While …
Offline-to-online Reinforcement Learning for Image-based Grasping with Scarce Demonstrations
B Chan, A Leung, J Bergstra - arXiv preprint arXiv:2410.14957, 2024 - arxiv.org
Offline-to-online reinforcement learning (O2O RL) aims to obtain a continually improving
policy as it interacts with the environment, while ensuring the initial behaviour is satisficing …
policy as it interacts with the environment, while ensuring the initial behaviour is satisficing …
Leveraging Domain-Unlabeled Data in Offline Reinforcement Learning across Two Domains
In this paper, we investigate an offline reinforcement learning (RL) problem where datasets
are collected from two domains. In this scenario, having datasets with domain labels …
are collected from two domains. In this scenario, having datasets with domain labels …
Value-Aided Conditional Supervised Learning for Offline RL
Offline reinforcement learning (RL) has seen notable advancements through return-
conditioned supervised learning (RCSL) and value-based methods, yet each approach …
conditioned supervised learning (RCSL) and value-based methods, yet each approach …
Optimistic Critic Reconstruction and Constrained Fine-Tuning for General Offline-to-Online RL
QW Luo, MK Xie, YW Wang, SJ Huang - The Thirty-eighth Annual … - openreview.net
Offline-to-online (O2O) reinforcement learning (RL) provides an effective means of
leveraging an offline pre-trained policy as initialization to improve performance rapidly with …
leveraging an offline pre-trained policy as initialization to improve performance rapidly with …