English
Related papers

Related papers: Depth creates no more spurious local minima

200 papers

Neural networks provide a rich class of high-dimensional, non-convex optimization problems. Despite their non-convexity, gradient-descent methods often successfully optimize these models. This has motivated a recent spur in research…

Optimization and Control · Mathematics 2020-06-18 Luca Venturi , Afonso S. Bandeira , Joan Bruna

There has been a lot of recent interest in trying to characterize the error surface of deep models. This stems from a long standing question. Given that deep networks are highly nonlinear systems optimized by local gradient methods, why do…

Machine Learning · Statistics 2017-02-20 Grzegorz Swirszcz , Wojciech Marian Czarnecki , Razvan Pascanu

Does over-parameterization eliminate sub-optimal local minima for neural networks? An affirmative answer was given by a classical result in [59] for 1-hidden-layer wide neural networks. A few recent works have extended the setting to…

Machine Learning · Computer Science 2020-11-17 Tian Ding , Dawei Li , Ruoyu Sun

One of the main difficulties in analyzing neural networks is the non-convexity of the loss function which may have many bad local minima. In this paper, we study the landscape of neural networks for binary classification tasks. Under mild…

Machine Learning · Statistics 2018-05-23 Shiyu Liang , Ruoyu Sun , Jason D. Lee , R. Srikant

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

Traditional landscape analysis of deep neural networks aims to show that no sub-optimal local minima exist in some appropriate sense. From this, one may be tempted to conclude that descent algorithms which escape saddle points will reach a…

Machine Learning · Computer Science 2020-01-01 Shiyu Liang , Ruoyu Sun , R. Srikant

By using the viewpoint of modern computational algebraic geometry, we explore properties of the optimization landscapes of the deep linear neural network models. After clarifying on the various definitions of "flat" minima, we show that the…

Machine Learning · Statistics 2018-10-19 Dhagash Mehta , Tianran Chen , Tingting Tang , Jonathan D. Hauenstein

This work finds the analytical expression of the global minima of a deep linear network with weight decay and stochastic neurons, a fundamental model for understanding the landscape of neural networks. Our result implies that the origin is…

Machine Learning · Statistics 2023-06-14 Liu Ziyin , Botao Li , Xiangming Meng

In this paper, we theoretically prove that gradient descent can find a global minimum of non-convex optimization of all layers for nonlinear deep neural networks of sizes commonly encountered in practice. The theory developed in this paper…

Machine Learning · Statistics 2020-06-18 Kenji Kawaguchi , Jiaoyang Huang

A main puzzle of deep networks revolves around the absence of overfitting despite large overparametrization and despite the large capacity demonstrated by zero training error on randomly labeled data. In this note, we show that the dynamics…

A main puzzle of deep neural networks (DNNs) revolves around the apparent absence of "overfitting", defined in this paper as follows: the expected error does not get worse when increasing the number of neurons or of iterations of gradient…

Machine Learning · Computer Science 2018-07-02 Tomaso Poggio , Qianli Liao , Brando Miranda , Andrzej Banburski , Xavier Boix , Jack Hidary

We examine the theoretical properties of enforcing priors provided by generative deep neural networks via empirical risk minimization. In particular we consider two models, one in which the task is to invert a generative neural network…

Information Theory · Computer Science 2018-12-27 Paul Hand , Vladislav Voroninski

It is widely conjectured that the reason that training algorithms for neural networks are successful because all local minima lead to similar performance, for example, see (LeCun et al., 2015, Choromanska et al., 2015, Dauphin et al.,…

Machine Learning · Computer Science 2018-03-06 Shiyu Liang , Ruoyu Sun , Yixuan Li , R. Srikant

For one-hidden-layer ReLU networks, we prove that all differentiable local minima are global inside differentiable regions. We give the locations and losses of differentiable local minima, and show that these local minima can be isolated…

Machine Learning · Computer Science 2020-06-18 Bo Liu

For nonconvex optimization in machine learning, this article proves that every local minimum achieves the globally optimal value of the perturbable gradient basis model at any differentiable point. As a result, nonconvex machine learning is…

Machine Learning · Statistics 2019-11-19 Kenji Kawaguchi , Jiaoyang Huang , Leslie Pack Kaelbling

The residual network is now one of the most effective structures in deep learning, which utilizes the skip connections to ``guarantee" the performance will not get worse. However, the non-convexity of the neural network makes it unclear…

Machine Learning · Computer Science 2020-06-11 Lifu Wang , Bo Shen , Ning Zhao , Zhiyuan Zhang

Nonconvex matrix recovery is known to contain no spurious local minima under a restricted isometry property (RIP) with a sufficiently small RIP constant $\delta$. If $\delta$ is too large, however, then counterexamples containing spurious…

Machine Learning · Computer Science 2020-04-28 Richard Y. Zhang , Somayeh Sojoudi , Javad Lavaei

It is well known that (stochastic) gradient descent has an implicit bias towards flat minima. In deep neural network training, this mechanism serves to screen out minima. However, the precise effect that this has on the trained network is…

Machine Learning · Computer Science 2020-08-11 Rotem Mulayoff , Tomer Michaeli

For fixed training data and network parameters in the other layers the L1 loss of a ReLU neural network as a function of the first layer's parameters is a piece-wise affine function. We use the Deep ReLU Simplex algorithm to iteratively…

Machine Learning · Statistics 2021-05-07 Peter Hinz

In this paper, we theoretically prove that the deep ReLU neural networks do not lie in spurious local minima in the loss landscape under the Neural Tangent Kernel (NTK) regime, that is, in the gradient descent training dynamics of the deep…

Machine Learning · Statistics 2022-05-20 Tohru Nitta