中文
相关论文

相关论文: The Low-Rank Simplicity Bias in Deep Networks

200 篇论文

Over-parameterization and adaptive methods have played a crucial role in the success of deep learning in the last decade. The widespread use of over-parameterization has forced us to rethink generalization by bringing forth new phenomena,…

机器学习 · 统计学 2020-12-01 Vatsal Shah , Soumya Basu , Anastasios Kyrillidis , Sujay Sanghavi

Deep neural networks are widely used prediction algorithms whose performance often improves as the number of weights increases, leading to over-parametrization. We consider a two-layered neural network whose first layer is frozen while the…

机器学习 · 计算机科学 2023-04-10 Roman Worschech , Bernd Rosenow

Works on implicit regularization have studied gradient trajectories during the optimization process to explain why deep networks favor certain kinds of solutions over others. In deep linear networks, it has been shown that gradient descent…

机器学习 · 计算机科学 2023-06-02 Dan Zhao

An overarching goal in machine learning is to build a generalizable model with few samples. To this end, overparameterization has been the subject of immense interest to explain the generalization ability of deep nets even when the size of…

机器学习 · 计算机科学 2022-01-19 Yue Sun , Adhyyan Narang , Halil Ibrahim Gulluk , Samet Oymak , Maryam Fazel

Understanding generalization of overparametrized neural networks remains a fundamental challenge in machine learning. Most of the literature mostly studies generalization from an interpolation point of view, taking convergence of parameters…

机器学习 · 计算机科学 2025-06-13 Etienne Boursier , Nicolas Flammarion

It has been observed \citep{zhang2016understanding} that deep neural networks can memorize: they achieve 100\% accuracy on training data. Recent theoretical results explained such behavior in highly overparametrized regimes, where the…

机器学习 · 计算机科学 2019-09-27 Rong Ge , Runzhe Wang , Haoyu Zhao

Promising resolutions of the generalization puzzle observe that the actual number of parameters in a deep network is much smaller than naive estimates suggest. The renormalization group is a compelling example of a problem which has very…

机器学习 · 计算机科学 2020-12-08 Anita de Mello Koch , Ellen de Mello Koch , Robert de Mello Koch

Simplicity bias is the concerning tendency of deep networks to over-depend on simple, weakly predictive features, to the exclusion of stronger, more complex features. This is exacerbated in real-world applications by limited training data…

机器学习 · 计算机科学 2023-06-07 Rishabh Tiwari , Pradeep Shenoy

Neural networks often exhibit simplicity bias, favoring simpler features over more complex ones, even when both are equally predictive. We introduce a novel method called imbalanced label coupling to explore and extend this simplicity bias…

机器学习 · 计算机科学 2024-10-23 Zhehang Du

Training deep neural networks in low rank, i.e. with factorised layers, is of particular interest to the community: it offers efficiency over unfactorised training in terms of both memory consumption and training time. Prior work has…

机器学习 · 计算机科学 2022-09-28 Siddhartha Rao Kamalakara , Acyr Locatelli , Bharat Venkitesh , Jimmy Ba , Yarin Gal , Aidan N. Gomez

It is frequently observed that overparameterized neural networks generalize well. Regarding such phenomena, existing theoretical work mainly devotes to linear settings or fully-connected neural networks. This paper studies the learning…

机器学习 · 统计学 2023-08-17 Tian-Yi Zhou , Xiaoming Huo

It is common practice in deep learning to use overparameterized networks and train for as long as possible; there are numerous studies that show, both theoretically and empirically, that such practices surprisingly do not unduly harm the…

机器学习 · 计算机科学 2020-03-05 Leslie Rice , Eric Wong , J. Zico Kolter

The ability of deep neural networks to generalize well in the overparameterized regime has become a subject of significant research interest. We show that overparameterized autoencoders exhibit memorization, a form of inductive bias that…

计算机视觉与模式识别 · 计算机科学 2019-09-05 Adityanarayanan Radhakrishnan , Karren Yang , Mikhail Belkin , Caroline Uhler

In recent years, great progress has been made in a variety of application domains thanks to the development of increasingly deeper neural networks. Unfortunately, the huge number of units of these networks makes them expensive both…

计算机视觉与模式识别 · 计算机科学 2018-10-12 Jose M. Alvarez , Mathieu Salzmann

Despite Deep Learning's (DL) empirical success, our theoretical understanding of its efficacy remains limited. One notable paradox is that while conventional wisdom discourages perfect data fitting, deep neural networks are designed to do…

机器学习 · 计算机科学 2024-02-06 Oria Gruber , Haim Avron

Deep learning methods are known to generalize well from training to future data, even in an overparametrized regime, where they could easily overfit. One explanation for this phenomenon is that even when their *ambient dimensionality*,…

机器学习 · 计算机科学 2025-05-22 Hossein Zakerinia , Dorsa Ghobadi , Christoph H. Lampert

Deep neural networks often contain far more parameters than training examples, yet they still manage to generalize well in practice. Classical complexity measures such as VC-dimension or PAC-Bayes bounds usually become vacuous in this…

机器学习 · 计算机科学 2025-08-26 Aviral Dhingra

Neural networks typically generalize well when fitting the data perfectly, even though they are heavily overparameterized. Many factors have been pointed out as the reason for this phenomenon, including an implicit bias of stochastic…

机器学习 · 计算机科学 2025-02-04 Amit Peleg , Matthias Hein

A recent line of work has shown that an overparametrized neural network can perfectly fit the training data, an otherwise often intractable nonconvex optimization problem. For (fully-connected) shallow networks, in the best case scenario,…

机器学习 · 计算机科学 2019-10-30 Armin Eftekhari , ChaeHwan Song , Volkan Cevher

Overparameterization, the condition where models have more parameters than necessary to fit their training loss, is a crucial factor for the success of deep learning. However, the characteristics of the features learned by overparameterized…

机器学习 · 计算机科学 2024-07-02 Ahmet Cagri Duzgun , Samy Jelassi , Yuanzhi Li