English
Related papers

Related papers: Stabilizing Q Learning Via Soft Mellowmax Operator

200 papers

Q-learning with value function approximation may have the poor performance because of overestimation bias and imprecise estimate. Specifically, overestimation bias is from the maximum operator over noise estimate, which is exaggerated using…

Machine Learning · Computer Science 2020-06-15 Gang Chen

The maximum element of the vector output by the Softmax function approaches zero as the input vector size increases. Transformer-based language models rely on Softmax to compute attention scores, causing the attention distribution to…

Computation and Language · Computer Science 2025-02-03 Ken M. Nakanishi

Softmax function is widely used in artificial neural networks for multiclass classification, multilabel classification, attention mechanisms, etc. However, its efficacy is often questioned in literature. The log-softmax loss has been shown…

Machine Learning · Computer Science 2020-11-24 Kunal Banerjee , Vishak Prasad C , Rishi Raj Gupta , Karthik Vyas , Anushree H , Biswajit Mishra

Offline Reinforcement Learning (RL) struggles with distributional shifts, leading to the $Q$-value overestimation for out-of-distribution (OOD) actions. Existing methods address this issue by imposing constraints; however, they often become…

Machine Learning · Computer Science 2025-06-11 Qingmao Yao , Zhichao Lei , Tianyuan Chen , Ziyue Yuan , Xuefan Chen , Jianxiang Liu , Faguo Wu , Xiao Zhang

The softmax content-based attention mechanism has proven to be very beneficial in many applications of recurrent neural networks. Nevertheless it suffers from two major computational limitations. First, its computations for an attention…

Machine Learning · Computer Science 2016-09-20 Alexandre de Brébisson , Pascal Vincent

This paper proposes a new regularization technique for reinforcement learning (RL) towards making policy and value functions smooth and stable. RL is known for the instability of the learning process and the sensitivity of the acquired…

Robotics · Computer Science 2023-07-04 Taisuke Kobayashi

We consider the problem of global optimization of an unknown non-convex smooth function with zeroth-order feedback. In this setup, an algorithm is allowed to adaptively query the underlying function at different locations and receives noisy…

Machine Learning · Statistics 2018-03-26 Yining Wang , Sivaraman Balakrishnan , Aarti Singh

Large language models (LLMs) have made transformed changes for human society. One of the key computation in LLMs is the softmax unit. This operation is important in LLMs because it allows the model to generate a distribution over possible…

Machine Learning · Computer Science 2023-04-27 Yichuan Deng , Zhihang Li , Zhao Song

In deep reinforcement learning, estimating the value function to evaluate the quality of states and actions is essential. The value function is often trained using the least squares method, which implicitly assumes a Gaussian error…

Machine Learning · Computer Science 2024-03-28 Motoki Omura , Takayuki Osa , Yusuke Mukuta , Tatsuya Harada

Large language models rely on attention mechanisms with a softmax activation. Yet the dominance of softmax over alternatives (e.g., component-wise or linear) remains poorly understood, and many theoretical works have focused on the…

Machine Learning · Computer Science 2026-02-27 O. Duranthon , P. Marion , C. Boyer , B. Loureiro , L. Zdeborová

The linear programming (LP) approach has a long history in the theory of approximate dynamic programming. When it comes to computation, however, the LP approach often suffers from poor scalability. In this work, we introduce a relaxed…

Systems and Control · Electrical Eng. & Systems 2020-12-01 Andrea Martinelli , Matilde Gargiani , John Lygeros

This paper addresses the problem of learning optimal policies for satisfying signal temporal logic (STL) specifications by agents with unknown stochastic dynamics. The system is modeled as a Markov decision process, in which the states…

Systems and Control · Computer Science 2016-09-26 Derya Aksaray , Austin Jones , Zhaodan Kong , Mac Schwager , Calin Belta

Convergence of Q-learning has been the subject of extensive study for decades. Among the available techniques, the ordinary differential equation (ODE) method is particularly appealing as a general-purpose, off-the-shelf tool for…

Machine Learning · Computer Science 2026-05-12 Donghwan Lee , Hyunjun Na

Bimodal, stochastic environments present a challenge to typical Reinforcement Learning problems. This problem is one that is surprisingly common in real world applications, being particularly applicable to pricing problems. In this paper we…

Machine Learning · Computer Science 2023-07-04 E. Hurwitz , N. Peace , G. Cevora

An agent's ability to leverage past experience is critical for efficiently solving new tasks. Prior work has focused on using value function estimates to obtain zero-shot approximations for solutions to a new task. In soft Q-learning, we…

Machine Learning · Computer Science 2024-06-27 Jacob Adamczyk , Volodymyr Makarenko , Stas Tiomkin , Rahul V. Kulkarni

In recent years, the softmax model and its fast approximations have become the de-facto loss functions for deep neural networks when dealing with multi-class prediction. This loss has been extended to language modeling and recommendation,…

Machine Learning · Statistics 2019-09-19 Ugo Tanielian , Flavian Vasile

Stochastic approximation (SA) that involves multiple coupled sequences, known as multiple-sequence SA (MSSA), finds diverse applications in the fields of signal processing and machine learning. However, existing theoretical understandings…

Machine Learning · Computer Science 2024-10-18 Yue Huang , Zhaoxian Wu , Shiqian Ma , Qing Ling

Simulation-to-decision learning enables safe policy training in digital environments without risking real-world deployment, and has become essential in mission-critical domains such as supply chains and industrial systems. However,…

Machine Learning · Computer Science 2026-03-11 Hongyu Cao , Jinghan Zhang , Kunpeng Liu , Dongjie Wang , Feng Xia , Haifeng Chen , Xiaohua Hu , Yanjie Fu

Double Q-learning is a classical method for reducing overestimation bias, which is caused by taking maximum estimated values in the Bellman operation. Its variants in the deep Q-learning paradigm have shown great promise in producing…

Machine Learning · Computer Science 2022-01-17 Zhizhou Ren , Guangxiang Zhu , Hao Hu , Beining Han , Jianglun Chen , Chongjie Zhang

In offline reinforcement learning (RL) we have no opportunity to explore so we must make assumptions that the data is sufficient to guide picking a good policy, taking the form of assuming some coverage, realizability, Bellman completeness,…

Machine Learning · Computer Science 2023-11-14 Masatoshi Uehara , Nathan Kallus , Jason D. Lee , Wen Sun