Hold me tight! Influence of discriminative features on deep network boundaries

• URL: http://arxiv.org/abs/2002.06349v4
• Date: Thu, 15 Oct 2020 07:16:47 GMT
• Title: Hold me tight! Influence of discriminative features on deep network boundaries
• Authors: Guillermo Ortiz-Jimenez, Apostolos Modas, Seyed-Mohsen Moosavi-Dezfooli, Pascal Frossard
• Abstract summary: We propose a new perspective that relates dataset features to the distance of samples to the decision boundary. This enables us to carefully tweak the position of the training samples and measure the induced changes on the boundaries of CNNs trained on large-scale vision datasets.
• Score: 63.627760598441796
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Important insights towards the explainability of neural networks reside in the characteristics of their decision boundaries. In this work, we borrow tools from the field of adversarial robustness, and propose a new perspective that relates dataset features to the distance of samples to the decision boundary. This enables us to carefully tweak the position of the training samples and measure the induced changes on the boundaries of CNNs trained on large-scale vision datasets. We use this framework to reveal some intriguing properties of CNNs. Specifically, we rigorously confirm that neural networks exhibit a high invariance to non-discriminative features, and show that the decision boundaries of a DNN can only exist as long as the classifier is trained with some features that hold them together. Finally, we show that the construction of the decision boundary is extremely sensitive to small perturbations of the training samples, and that changes in certain directions can lead to sudden invariances in the orthogonal ones. This is precisely the mechanism that adversarial training uses to achieve robustness.

Related papers

• Neural Networks Decoded: Targeted and Robust Analysis of Neural Network Decisions via Causal Explanations and Reasoning [9.947555560412397]
  We introduce TRACER, a novel method grounded in causal inference theory to estimate the causal dynamics underpinning DNN decisions. Our approach systematically intervenes on input features to observe how specific changes propagate through the network, affecting internal activations and final outputs. TRACER further enhances explainability by generating counterfactuals that reveal possible model biases and offer contrastive explanations for misclassifications.
  arXiv  Detail & Related papers  (2024-10-07T20:44:53Z)
• Application of Tensorized Neural Networks for Cloud Classification [0.0]
  Convolutional neural networks (CNNs) have gained widespread usage across various fields such as weather forecasting, computer vision, autonomous driving, and medical image analysis. However, the practical implementation and commercialization of CNNs in these domains are hindered by challenges related to model sizes, overfitting, and computational time. We propose a groundbreaking approach that involves tensorizing the dense layers in the CNN to reduce model size and computational time.
  arXiv  Detail & Related papers  (2024-03-21T06:28:22Z)
• Why do Learning Rates Transfer? Reconciling Optimization and Scaling Limits for Deep Learning [77.82908213345864]
  We find empirical evidence that learning rate transfer can be attributed to the fact that under $mu$P and its depth extension, the largest eigenvalue of the training loss Hessian is largely independent of the width and depth of the network. We show that under the neural tangent kernel (NTK) regime, the sharpness exhibits very different dynamics at different scales, thus preventing learning rate transfer.
  arXiv  Detail & Related papers  (2024-02-27T12:28:01Z)
• DARE: Towards Robust Text Explanations in Biomedical and Healthcare Applications [54.93807822347193]
  We show how to adapt attribution robustness estimation methods to a given domain, so as to take into account domain-specific plausibility. Next, we provide two methods, adversarial training and FAR training, to mitigate the brittleness characterized by DARE. Finally, we empirically validate our methods with extensive experiments on three established biomedical benchmarks.
  arXiv  Detail & Related papers  (2023-07-05T08:11:40Z)
• Feature-Learning Networks Are Consistent Across Widths At Realistic Scales [72.27228085606147]
  We study the effect of width on the dynamics of feature-learning neural networks across a variety of architectures and datasets. Early in training, wide neural networks trained on online data have not only identical loss curves but also agree in their point-wise test predictions throughout training. We observe, however, that ensembles of narrower networks perform worse than a single wide network.
  arXiv  Detail & Related papers  (2023-05-28T17:09:32Z)
• Unfolding Local Growth Rate Estimates for (Almost) Perfect Adversarial Detection [22.99930028876662]
  Convolutional neural networks (CNN) define the state-of-the-art solution on many perceptual tasks. Current CNN approaches largely remain vulnerable against adversarial perturbations of the input that have been crafted specifically to fool the system. We propose a simple and light-weight detector, which leverages recent findings on the relation between networks' local intrinsic dimensionality (LID) and adversarial attacks.
  arXiv  Detail & Related papers  (2022-12-13T17:51:32Z)
• Robustness Threats of Differential Privacy [70.818129585404]
  We experimentally demonstrate that networks, trained with differential privacy, in some settings might be even more vulnerable in comparison to non-private versions. We study how the main ingredients of differentially private neural networks training, such as gradient clipping and noise addition, affect the robustness of the model.
  arXiv  Detail & Related papers  (2020-12-14T18:59:24Z)
• Attribute-Guided Adversarial Training for Robustness to Natural Perturbations [64.35805267250682]
  We propose an adversarial training approach which learns to generate new samples so as to maximize exposure of the classifier to the attributes-space. Our approach enables deep neural networks to be robust against a wide range of naturally occurring perturbations.
  arXiv  Detail & Related papers  (2020-12-03T10:17:30Z)
• Understanding the Decision Boundary of Deep Neural Networks: An Empirical Study [0.4499833362998487]
  We study the minimum distance of data points to the decision boundary and how this margin evolves over the training of a deep neural network. We observe that the decision boundary moves closer to natural images over training. On the other hand, adversarial training appears to have the potential to prevent this undesired convergence of the decision boundary.
  arXiv  Detail & Related papers  (2020-02-05T14:34:22Z)

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

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.