English
Related papers

Related papers: Local to Global: Learning Dynamics and Effect of I…

200 papers

Grokking, a delayed generalization in neural networks after perfect training performance, has been observed in Transformers and MLPs, but the components driving it remain underexplored. We show that embeddings are central to grokking:…

Machine Learning · Computer Science 2025-05-22 H. V. AlquBoj , Hilal AlQuabeh , Velibor Bojkovic , Munachiso Nwadike , Kentaro Inui

Transformers have demonstrated remarkable success across various applications. However, the success of transformers have not been understood in theory. In this work, we give a case study of how transformers can be trained to learn a classic…

Machine Learning · Statistics 2025-04-14 Chenyang Zhang , Xuran Meng , Yuan Cao

Can Transformers predict new syllogisms by composing established ones? More generally, what type of targets can be learned by such models from scratch? Recent works show that Transformers can be Turing-complete in terms of expressivity, but…

Machine Learning · Computer Science 2024-11-05 Emmanuel Abbe , Samy Bengio , Aryo Lotfi , Colin Sandon , Omid Saremi

Variable-length Markov chains on finite quivers provide a natural framework for context-dependent stochastic growth under incidence constraints. I study quiver-valued variable-length Markov chains observed through finite boundary windows…

Probability · Mathematics 2026-04-14 Oleg Kiriukhin

Grokking -- the delayed transition from memorization to generalization in small algorithmic tasks -- remains poorly understood. We present a geometric analysis of optimization dynamics in transformers trained on modular arithmetic. PCA of…

Machine Learning · Computer Science 2026-04-06 Yongzhong Xu

A significant advance in accelerating neural network training has been the development of normalization methods, permitting the training of deep models both faster and with better accuracy. These advances come with practical challenges: for…

Machine Learning · Computer Science 2019-03-05 Jasmine Collins , Johannes Balle , Jonathon Shlens

Spatiotemporal graph neural networks have shown to be effective in time series forecasting applications, achieving better performance than standard univariate predictors in several settings. These architectures take advantage of a graph…

Machine Learning · Computer Science 2023-11-13 Andrea Cini , Ivan Marisca , Daniele Zambon , Cesare Alippi

We consider the problem of test-time domain generalization, where a model is trained on several source domains and adjusted on target domains never seen during training. Different from the common methods that fine-tune the model or adjust…

Machine Learning · Computer Science 2025-02-19 Sameer Ambekar , Zehao Xiao , Xiantong Zhen , Cees G. M. Snoek

We theoretically characterize gradient descent dynamics in deep linear networks trained at large width from random initialization and on large quantities of random data. Our theory captures the ``wider is better" effect of…

Machine Learning · Computer Science 2025-06-17 Blake Bordelon , Cengiz Pehlevan

Neural networks trained via gradient descent with random initialization and without any regularization enjoy good generalization performance in practice despite being highly overparametrized. A promising direction to explain this phenomenon…

Machine Learning · Computer Science 2022-05-17 Hancheng Min , Salma Tarmoun , Rene Vidal , Enrique Mallada

Recently many first and second order variants of SGD have been proposed to facilitate training of Deep Neural Networks (DNNs). A common limitation of these works stem from the fact that they use the same learning rate across all instances…

Machine Learning · Computer Science 2021-05-31 Shreyas Saxena , Nidhi Vyas , Dennis DeCoste

Agents trained with deep reinforcement learning algorithms are capable of performing highly complex tasks including locomotion in continuous environments. We investigate transferring the learning acquired in one task to a set of previously…

Machine Learning · Computer Science 2024-03-06 Suzan Ece Ada , Emre Ugur , H. Levent Akin

Many types of neural network layers rely on matrix properties such as invertibility or orthogonality. Retaining such properties during optimization with gradient-based stochastic optimizers is a challenging task, which is usually addressed…

Machine Learning · Statistics 2020-12-02 Andreas Krämer , Jonas Köhler , Frank Noé

In this paper, we consider the supervised pre-trained transformer for a class of sequential decision-making problems. The class of considered problems is a subset of the general formulation of reinforcement learning in that there is no…

Machine Learning · Computer Science 2024-10-03 Hanzhao Wang , Yu Pan , Fupeng Sun , Shang Liu , Kalyan Talluri , Guanting Chen , Xiaocheng Li

Recent results suggest that reinitializing a subset of the parameters of a neural network during training can improve generalization, particularly for small training sets. We study the impact of different reinitialization methods in several…

Machine Learning · Computer Science 2021-09-02 Ibrahim Alabdulmohsin , Hartmut Maennel , Daniel Keysers

We analyze the generalization gap (gap between the training and test errors) when training a potentially over-parametrized model using a Markovian stochastic training algorithm, initialized from some distribution $\theta_0 \sim p_0$. We…

Machine Learning · Computer Science 2025-10-21 Itamar Harel , Yonathan Wolanowsky , Gal Vardi , Nathan Srebro , Daniel Soudry

Learning reduced descriptions of chaotic many-body dynamics is fundamentally challenging: although microscopic equations are Markovian, collective observables exhibit strong memory and exponential sensitivity to initial conditions and…

Computational Physics · Physics 2026-01-28 Ho Jang , Gia-Wei Chern

Learning with identical train and test distributions has been extensively investigated both practically and theoretically. Much remains to be understood, however, in statistical learning under distribution shifts. This paper focuses on a…

Machine Learning · Computer Science 2024-11-01 Omar Montasser , Han Shao , Emmanuel Abbe

Locally supervised learning aims to train a neural network based on a local estimation of the global loss function at each decoupled module of the network. Auxiliary networks are typically appended to the modules to approximate the gradient…

Machine Learning · Computer Science 2022-08-02 Hasnain Irshad Bhatti , Jaekyun Moon

We propose an end-to-end differentiable training paradigm for stable training of a rationalized transformer classifier. Our approach results in a single model that simultaneously classifies a sample and scores input tokens based on their…

Computation and Language · Computer Science 2025-08-18 Marc Brinner , Sina Zarrieß