English
Related papers

Related papers: Hidden Minima in Two-Layer ReLU Networks

200 papers

We study the problem of PAC learning a linear combination of $k$ ReLU activations under the standard Gaussian distribution on $\mathbb{R}^d$ with respect to the square loss. Our main result is an efficient algorithm for this learning task…

Machine Learning · Computer Science 2023-07-26 Ilias Diakonikolas , Daniel M. Kane

Many modern neural network architectures are trained in an overparameterized regime where the parameters of the model exceed the size of the training dataset. Sufficiently overparameterized neural network architectures in principle have the…

Machine Learning · Computer Science 2019-02-14 Samet Oymak , Mahdi Soltanolkotabi

Deep learning requires several design choices, such as the nodes' activation functions and the widths, types, and arrangements of the layers. One consideration when making these choices is the vanishing-gradient problem, which is the…

Machine Learning · Computer Science 2021-06-07 Leni Ven , Johannes Lederer

In this work, we investigate a particular implicit bias in gradient descent training, which we term "Feature Averaging," and argue that it is one of the principal factors contributing to the non-robustness of deep neural networks. We show…

Machine Learning · Computer Science 2025-03-04 Binghui Li , Zhixuan Pan , Kaifeng Lyu , Jian Li

Despite the empirical success of using Adversarial Training to defend deep learning models against adversarial perturbations, so far, it still remains rather unclear what the principles are behind the existence of adversarial perturbations,…

Machine Learning · Computer Science 2022-06-14 Zeyuan Allen-Zhu , Yuanzhi Li

Recent results in the literature indicate that a residual network (ResNet) composed of a single residual block outperforms linear predictors, in the sense that all local minima in its optimization landscape are at least as good as the best…

Machine Learning · Computer Science 2019-10-30 Chulhee Yun , Suvrit Sra , Ali Jadbabaie

Giving provable guarantees for learning neural networks is a core challenge of machine learning theory. Most prior work gives parameter recovery guarantees for one hidden layer networks, however, the networks used in practice have multiple…

Machine Learning · Computer Science 2020-02-21 Surbhi Goel , Rina Panigrahy

Under mild assumptions, we investigate the geometry of the loss landscape for two-layer neural networks in the vicinity of global minima. Utilizing novel techniques, we demonstrate: (i) how global minima with zero generalization error…

Machine Learning · Computer Science 2025-04-11 Leyang Zhang , Yaoyu Zhang , Tao Luo

Over-parameterized neural networks generalize well in practice without any explicit regularization. Although it has not been proven yet, empirical evidence suggests that implicit regularization plays a crucial role in deep learning and…

Machine Learning · Computer Science 2019-03-07 Masayoshi Kubo , Ryotaro Banno , Hidetaka Manabe , Masataka Minoji

Characterization of local minima draws much attention in theoretical studies of deep learning. In this study, we investigate the distribution of parameters in an over-parametrized finite neural network trained by ridge regularized empirical…

Machine Learning · Computer Science 2021-02-22 Sho Sonoda , Isao Ishikawa , Masahiro Ikeda

To investigate the theoretical foundations of deep learning from the viewpoint of the minimum description length (MDL) principle, we analyse risk bounds of MDL estimators based on two-stage codes for simple two-layers neural networks (NNs)…

Information Theory · Computer Science 2024-11-19 Yoshinari Takeishi , Jun'ichi Takeuchi

Training deep neural networks is a challenging non-convex optimization problem. Recent work has proven that the strong duality holds (which means zero duality gap) for regularized finite-width two-layer ReLU networks and consequently…

Machine Learning · Computer Science 2023-03-08 Yifei Wang , Tolga Ergen , Mert Pilanci

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…

Machine Learning · Computer Science 2025-06-13 Etienne Boursier , Nicolas Flammarion

We prove several hardness results for training depth-2 neural networks with the ReLU activation function; these networks are simply weighted sums (that may include negative coefficients) of ReLUs. Our goal is to output a depth-2 neural…

Machine Learning · Computer Science 2020-11-30 Surbhi Goel , Adam Klivans , Pasin Manurangsi , Daniel Reichman

This paper provides a least squares formulation for the training of a 2-layer convolutional neural network using quadratic activation functions, a 2-norm loss function, and no regularization term. Using this method, an analytic expression…

Machine Learning · Computer Science 2024-11-18 Zachary Yetman Van Egmond , Luis Rodrigues

Rectified Linear Units (ReLU) have become the main model for the neural units in current deep learning systems. This choice has been originally suggested as a way to compensate for the so called vanishing gradient problem which can undercut…

Disordered Systems and Neural Networks · Physics 2024-05-06 Carlo Baldassi , Enrico M. Malatesta , Riccardo Zecchina

It is well-known that the Hessian of deep loss landscape matters to optimization, generalization, and even robustness of deep learning. Recent works empirically discovered that the Hessian spectrum in deep learning has a two-component…

Machine Learning · Computer Science 2022-08-02 Zeke Xie , Qian-Yuan Tang , Yunfeng Cai , Mingming Sun , Ping Li

In this paper, we study the optimality gap between two-layer ReLU networks regularized with weight decay and their convex relaxations. We show that when the training data is random, the relative optimality gap between the original problem…

Machine Learning · Computer Science 2024-07-15 Sungyoon Kim , Mert Pilanci

We study feature learning in two-layer neural networks within the linear-width regime, where the number of hidden neurons, sample size, and input dimension scale proportionally. While recent work has analyzed feature learning via a single…

Machine Learning · Statistics 2026-05-25 Behrad Moniri , Hamed Hassani

We present a simple proof for the benefit of depth in multi-layer feedforward network with rectified activation ("depth separation"). Specifically we present a sequence of classification problems indexed by $m$ such that (a) for any fixed…

Machine Learning · Computer Science 2021-01-19 Asaf Amrami , Yoav Goldberg