Related papers: The Double-Edged Sword of Implicit Bias: Generalization vs. Robustness in ReLU Networks

The Double-Edged Sword of Implicit Bias: Generalization vs. Robustness in ReLU Networks

URL: http://arxiv.org/abs/2303.01456v2
Date: Tue, 31 Oct 2023 09:23:27 GMT
Title: The Double-Edged Sword of Implicit Bias: Generalization vs. Robustness in ReLU Networks
Authors: Spencer Frei and Gal Vardi and Peter L. Bartlett and Nathan Srebro
Abstract summary: We study the implications of the implicit bias of gradient flow on generalization and adversarial robustness in ReLU networks. In two-layer ReLU networks gradient flow is biased towards solutions that generalize well, but are highly vulnerable to adversarial examples.
Score: 64.12052498909105
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Abstract: In this work, we study the implications of the implicit bias of gradient flow on generalization and adversarial robustness in ReLU networks. We focus on a setting where the data consists of clusters and the correlations between cluster means are small, and show that in two-layer ReLU networks gradient flow is biased towards solutions that generalize well, but are highly vulnerable to adversarial examples. Our results hold even in cases where the network has many more parameters than training examples. Despite the potential for harmful overfitting in such overparameterized settings, we prove that the implicit bias of gradient flow prevents it. However, the implicit bias also leads to non-robust solutions (susceptible to small adversarial $\ell_2$-perturbations), even though robust networks that fit the data exist.

Related papers

Benign Overfitting for Regression with Trained Two-Layer ReLU Networks [14.36840959836957]
We study the least-square regression problem with a two-layer fully-connected neural network, with ReLU activation function, trained by gradient flow. Our first result is a generalization result, that requires no assumptions on the underlying regression function or the noise other than that they are bounded.
arXiv Detail & Related papers (2024-10-08T16:54:23Z)
Can Implicit Bias Imply Adversarial Robustness? [36.467655284354606]
The implicit bias of gradient-based training algorithms has been considered mostly beneficial as it leads to trained networks that often generalize well. However, Frei et al. (2023) show that such implicit bias can harm adversarial robustness. Our results highlight the importance of the interplay between data structure and architecture in the implicit bias and robustness of trained networks.
arXiv Detail & Related papers (2024-05-24T21:09:53Z)
GFlowOut: Dropout with Generative Flow Networks [76.59535235717631]
Monte Carlo Dropout has been widely used as a relatively cheap way for approximate Inference. Recent works show that the dropout mask can be viewed as a latent variable, which can be inferred with variational inference. GFlowOutleverages the recently proposed probabilistic framework of Generative Flow Networks (GFlowNets) to learn the posterior distribution over dropout masks.
arXiv Detail & Related papers (2022-10-24T03:00:01Z)
Self-supervised debiasing using low rank regularization [59.84695042540525]
Spurious correlations can cause strong biases in deep neural networks, impairing generalization ability. We propose a self-supervised debiasing framework potentially compatible with unlabeled samples. Remarkably, the proposed debiasing framework significantly improves the generalization performance of self-supervised learning baselines.
arXiv Detail & Related papers (2022-10-11T08:26:19Z)
On the Effective Number of Linear Regions in Shallow Univariate ReLU Networks: Convergence Guarantees and Implicit Bias [50.84569563188485]
We show that gradient flow converges in direction when labels are determined by the sign of a target network with $r$ neurons. Our result may already hold for mild over- parameterization, where the width is $tildemathcalO(r)$ and independent of the sample size.
arXiv Detail & Related papers (2022-05-18T16:57:10Z)
Gradient Methods Provably Converge to Non-Robust Networks [40.83290846983707]
In adversarial networks, depth-$2LU$ Reperturbible gradient networks are provably non-robust. We show that the well-known implicit bias towards a margin induces bias towards non-robust networks.
arXiv Detail & Related papers (2022-02-09T08:58:54Z)
Implicit Regularization Towards Rank Minimization in ReLU Networks [34.41953136999683]
We study the conjectured relationship between the implicit regularization in neural networks and rank minimization. We focus on nonlinear ReLU networks, providing several new positive and negative results.
arXiv Detail & Related papers (2022-01-30T09:15:44Z)
Predicting Deep Neural Network Generalization with Perturbation Response Curves [58.8755389068888]
We propose a new framework for evaluating the generalization capabilities of trained networks. Specifically, we introduce two new measures for accurately predicting generalization gaps. We attain better predictive scores than the current state-of-the-art measures on a majority of tasks in the Predicting Generalization in Deep Learning (PGDL) NeurIPS 2020 competition.
arXiv Detail & Related papers (2021-06-09T01:37:36Z)
Unbiased Risk Estimators Can Mislead: A Case Study of Learning with Complementary Labels [92.98756432746482]
We study a weakly supervised problem called learning with complementary labels. We show that the quality of gradient estimation matters more in risk minimization. We propose a novel surrogate complementary loss(SCL) framework that trades zero bias with reduced variance.
arXiv Detail & Related papers (2020-07-05T04:19:37Z)
Towards More Practical Adversarial Attacks on Graph Neural Networks [14.78539966828287]
We study the black-box attacks on graph neural networks (GNNs) under a novel and realistic constraint. We show that the structural inductive biases of GNN models can be an effective source for this type of attacks.
arXiv Detail & Related papers (2020-06-09T05:27:39Z)

This list is automatically generated from the titles and abstracts of the papers in this site.