English
Related papers

Related papers: Biased Gradient Estimate with Drastic Variance Red…

200 papers

The concept of the value-gradient is introduced and developed in the context of reinforcement learning. It is shown that by learning the value-gradients exploration or stochastic behaviour is no longer needed to find locally optimal…

Neural and Evolutionary Computing · Computer Science 2008-03-26 Michael Fairbank

A significant challenge for the practical application of reinforcement learning in the real world is the need to specify an oracle reward function that correctly defines a task. Inverse reinforcement learning (IRL) seeks to avoid this…

Machine Learning · Computer Science 2019-10-16 Kelvin Xu , Ellis Ratner , Anca Dragan , Sergey Levine , Chelsea Finn

Likelihood-to-evidence ratio estimation is usually cast as either a binary (NRE-A) or a multiclass (NRE-B) classification task. In contrast to the binary classification framework, the current formulation of the multiclass version has an…

Machine Learning · Statistics 2024-07-08 Benjamin Kurt Miller , Christoph Weniger , Patrick Forré

We show that deep reinforcement learning algorithms can retain their ability to learn without resetting network parameters in settings where the number of gradient updates greatly exceeds the number of environment samples by combatting…

Machine Learning · Computer Science 2024-08-06 Marcel Hussing , Claas Voelcker , Igor Gilitschenski , Amir-massoud Farahmand , Eric Eaton

Gradient-based learning in multi-agent systems is difficult because the gradient derives from a first-order model which does not account for the interaction between agents' learning processes. LOLA (arXiv:1709.04326) accounts for this by…

Machine Learning · Computer Science 2023-12-12 Tim Cooijmans , Milad Aghajohari , Aaron Courville

Modern meta-reinforcement learning (Meta-RL) methods are mainly developed based on model-agnostic meta-learning, which performs policy gradient steps across tasks to maximize policy performance. However, the gradient conflict problem is…

Artificial Intelligence · Computer Science 2022-09-22 Haozhi Wang , Qing Wang , Yunfeng Shao , Dong Li , Jianye Hao , Yinchuan Li

Reward design remains a significant bottleneck in applying reinforcement learning (RL) to real-world problems. A popular alternative is reward learning, where reward functions are inferred from human feedback rather than manually specified.…

Machine Learning · Computer Science 2026-01-16 Chaitanya Kharyal , Calarina Muslimani , Matthew E. Taylor

We study the connection between gradient-based meta-learning and convex op-timisation. We observe that gradient descent with momentum is a special case of meta-gradients, and building on recent results in optimisation, we prove convergence…

Machine Learning · Computer Science 2023-01-10 Sebastian Flennerhag , Tom Zahavy , Brendan O'Donoghue , Hado van Hasselt , András György , Satinder Singh

In model-free deep reinforcement learning (RL) algorithms, using noisy value estimates to supervise policy evaluation and optimization is detrimental to the sample efficiency. As this noise is heteroscedastic, its effects can be mitigated…

Machine Learning · Computer Science 2022-05-04 Vincent Mai , Kaustubh Mani , Liam Paull

Several approximate inference algorithms have been proposed to minimize an alpha-divergence between an approximating distribution and a target distribution. Many of these algorithms introduce bias, the magnitude of which becomes problematic…

Machine Learning · Statistics 2021-10-27 Tomas Geffner , Justin Domke

In natural language processing tasks, pure reinforcement learning (RL) fine-tuning methods often suffer from inefficient exploration and slow convergence; while supervised fine-tuning (SFT) methods, although efficient in training, have…

Computation and Language · Computer Science 2025-09-17 Min Zeng , Jingfei Sun , Xueyou Luo , Caiquan Liu , Shiqi Zhang , Li Xie , Xiaoxin Chen

Reinforcement Learning with Verifiable Rewards (RLVR) reliably improves the reasoning performance of large language models, yet it appears to modify only a small fraction of parameters. We revisit this paradox and show that sparsity is a…

Gradient-based methods enable efficient search capabilities in high dimensions. However, in order to apply them effectively in offline optimization paradigms such as offline Reinforcement Learning (RL) or Imitation Learning (IL), we require…

Machine Learning · Computer Science 2023-10-18 H. J. Terry Suh , Glen Chou , Hongkai Dai , Lujie Yang , Abhishek Gupta , Russ Tedrake

Policy-gradient methods are widely used in reinforcement learning, yet training often becomes unstable or slows down as learning progresses. We study this phenomenon through the noise-to-signal ratio (NSR) of a policy-gradient estimator,…

Optimization and Control · Mathematics 2026-02-10 Haoyu Han , Heng Yang

Stochastic gradient methods are among the most important algorithms in training machine learning problems. While classical assumptions such as strong convexity allow a simple analysis they are rarely satisfied in applications. In recent…

Machine Learning · Computer Science 2025-03-18 Simon Weissmann , Sara Klein , Waïss Azizian , Leif Döring

Learning to rank (LTR) plays a crucial role in various Information Retrieval (IR) tasks. Although supervised LTR methods based on fine-grained relevance labels (e.g., document-level annotations) have achieved significant success, their…

Information Retrieval · Computer Science 2025-08-21 Yiteng Tu , Zhichao Xu , Tao Yang , Weihang Su , Yujia Zhou , Yiqun Liu , Fen Lin , Qin Liu , Qingyao Ai

Although outcome-based reinforcement learning (RL) significantly advances the mathematical reasoning capabilities of Large Language Models (LLMs), its reliance on computationally expensive ground-truth annotations imposes a severe…

Machine Learning · Computer Science 2026-03-18 Zelin Zhang , Fei Cheng , Chenhui Chu

Gradient-based optimization has been a cornerstone of machine learning that enabled the vast advances of Artificial Intelligence (AI) development over the past decades. However, this type of optimization requires differentiation, and with…

Meta-reinforcement learning (meta-RL) aims to learn from multiple training tasks the ability to adapt efficiently to unseen test tasks. Despite the success, existing meta-RL algorithms are known to be sensitive to the task distribution…

Machine Learning · Computer Science 2021-03-02 Zichuan Lin , Garrett Thomas , Guangwen Yang , Tengyu Ma

In Reinforcement Learning (RL), agents aim at maximizing cumulative rewards in a given environment. During the learning process, RL agents face the dilemma of exploitation and exploration: leveraging existing knowledge to acquire rewards or…

Machine Learning · Computer Science 2023-10-24 Chenfan Weng , Zhongguo Li