English
Related papers

Related papers: AdamO: A Collapse-Suppressed Optimizer for Offline…

200 papers

Offline reinforcement learning (RL) enables policy learning from pre-collected offline datasets, relaxing the need to interact directly with the environment. However, limited by the quality of offline datasets, it generally fails to learn…

Machine Learning · Computer Science 2025-09-03 Xingshuai Huang , Di Wu , Benoit Boulet

We prove local convergence of several notable gradient descent algorithms used in machine learning, for which standard stochastic gradient descent theory does not apply directly. This includes, first, online algorithms for recurrent models…

Dynamical Systems · Mathematics 2021-01-11 Pierre-Yves Massé , Yann Ollivier

Model-based offline reinforcement learning (RL) has made remarkable progress, offering a promising avenue for improving generalization with synthetic model rollouts. Existing works primarily focus on incorporating pessimism for policy…

Machine Learning · Computer Science 2024-01-12 Yuanzhao Zhai , Yiying Li , Zijian Gao , Xudong Gong , Kele Xu , Dawei Feng , Ding Bo , Huaimin Wang

Offline reinforcement learning (RL) aims to find an optimal policy for Markov decision processes (MDPs) using a pre-collected dataset. In this work, we revisit the linear programming (LP) reformulation of Markov decision processes for…

Machine Learning · Computer Science 2024-12-11 Asuman Ozdaglar , Sarath Pattathil , Jiawei Zhang , Kaiqing Zhang

Offline reinforcement learning (RL) refers to the problem of learning policies from a static dataset of environment interactions. Offline RL enables extensive use and re-use of historical datasets, while also alleviating safety concerns…

Machine Learning · Computer Science 2020-12-22 Rafael Rafailov , Tianhe Yu , Aravind Rajeswaran , Chelsea Finn

We focus on the task of approximating the optimal value function in deep reinforcement learning. This iterative process is comprised of solving a sequence of optimization problems where the loss function changes per iteration. The common…

Machine Learning · Computer Science 2023-11-16 Kavosh Asadi , Rasool Fakoor , Shoham Sabach

Adaptive optimizers with decoupled weight decay, such as AdamW, are the de facto standard for pre-training large transformer-based generative models. Yet the quadratic nature of the $\ell_2$ penalty embedded in weight decay drives all…

Machine Learning · Computer Science 2025-11-19 Fu-Ming Guo , Yingfang Fan

The Adam optimizer is a cornerstone of modern deep learning, yet the empirical necessity of each of its individual components is often taken for granted. This paper presents a focused investigation into the role of bias-correction, a…

Machine Learning · Computer Science 2025-11-27 Sam Laing , Antonio Orvieto

Continual learning has emerged as a pivotal area of research, primarily due to its advantageous characteristic that allows models to persistently acquire and retain information. However, catastrophic forgetting can severely impair model…

Machine Learning · Computer Science 2026-01-28 KaiHui Huang , RunQing Wu , JinHui Sheng , HanYi Zhang , Ling Ge , JinYu Guo , Fei Ye

Offline reinforcement learning (RL) holds promise as a means to learn high-reward policies from a static dataset, without the need for further environment interactions. However, a key challenge in offline RL lies in effectively stitching…

Machine Learning · Computer Science 2023-09-14 Siddarth Venkatraman , Shivesh Khaitan , Ravi Tej Akella , John Dolan , Jeff Schneider , Glen Berseth

Offline pretraining with a static dataset followed by online fine-tuning (offline-to-online, or OtO) is a paradigm well matched to a real-world RL deployment process. In this scenario, we aim to find the best-performing policy within a…

Machine Learning · Computer Science 2024-06-24 Trevor McInroe , Adam Jelley , Stefano V. Albrecht , Amos Storkey

Task-oriented dialogue systems aim to fulfill user goals through natural language interactions. They are ideally evaluated with human users, which however is unattainable to do at every iteration of the development phase. Simulated users…

Computation and Language · Computer Science 2022-09-05 Nurul Lubis , Christian Geishauser , Hsien-Chin Lin , Carel van Niekerk , Michael Heck , Shutong Feng , Milica Gašić

Off-policy reinforcement learning (RL) from pixel observations is notoriously unstable. As a result, many successful algorithms must combine different domain-specific practices and auxiliary losses to learn meaningful behaviors in complex…

Machine Learning · Computer Science 2022-07-05 Edoardo Cetin , Philip J. Ball , Steve Roberts , Oya Celiktutan

Offline reinforcement learning aims to learn from pre-collected datasets without active exploration. This problem faces significant challenges, including limited data availability and distributional shifts. Existing approaches adopt a…

Machine Learning · Computer Science 2024-10-01 Yue Wang , Jinjun Xiong , Shaofeng Zou

In this paper, we focus on a theory-practice gap for Adam and its variants (AMSgrad, AdamNC, etc.). In practice, these algorithms are used with a constant first-order moment parameter $\beta_{1}$ (typically between $0.9$ and $0.99$). In…

Machine Learning · Statistics 2020-03-24 Ahmet Alacaoglu , Yura Malitsky , Panayotis Mertikopoulos , Volkan Cevher

In order to obviate the requirement of drift dynamics in adaptive dynamic programming (ADP), integral reinforcement learning (IRL) has been proposed as an alternate formulation of Bellman equation.However control coupling dynamics is still…

Systems and Control · Electrical Eng. & Systems 2020-05-11 Amardeep Mishra , Satadal Ghosh

Reinforcement learning has been central to recent advances in large language model reasoning, but most algorithms rely on on-policy training that demands fresh rollouts at every update, limiting efficiency and scalability. Asynchronous RL…

Machine Learning · Computer Science 2025-10-29 Haizhong Zheng , Jiawei Zhao , Beidi Chen

We consider offline policy optimization (OPO) in contextual bandits, where one is given a fixed dataset of logged interactions. While pessimistic regularizers are typically used to mitigate distribution shift, prior implementations thereof…

Machine Learning · Computer Science 2023-10-27 Lequn Wang , Akshay Krishnamurthy , Aleksandrs Slivkins

Single-domain offline reinforcement learning (RL) often suffers from limited data coverage, while cross-domain offline RL handles this issue by leveraging additional data from other domains with dynamics shifts. However, existing studies…

Machine Learning · Computer Science 2026-03-10 Zhongjian Qiao , Rui Yang , Jiafei Lyu , Xiu Li , Zhongxiang Dai , Zhuoran Yang , Siyang Gao , Shuang Qiu

Reinforcement learning (RL) in large environments often suffers from severe computational bottlenecks, as conventional regret minimization algorithms require repeated, costly calls to planning and statistical estimation oracles. While…

Machine Learning · Computer Science 2026-05-04 Haichen Hu , Jian Qian , David Simchi-Levi