Understanding Parameter Saliency via Extreme Value Theory

• URL: http://arxiv.org/abs/2310.17951v2
• Date: Tue, 5 Dec 2023 05:00:20 GMT
• Title: Understanding Parameter Saliency via Extreme Value Theory
• Abstract summary: The concept of parameter saliency is proposed and used to diagnose convolutional neural networks (CNNs) It is shown that fine-tuning the top ranking salient filters efficiently corrects misidentification on ImageNet. However, there is still a knowledge gap in terms of understanding why parameter saliency ranking can find the filters inducing misidentification.
• Score: 44.81687972865167
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Deep neural networks are being increasingly implemented throughout society in recent years. It is useful to identify which parameters trigger misclassification in diagnosing undesirable model behaviors. The concept of parameter saliency is proposed and used to diagnose convolutional neural networks (CNNs) by ranking convolution filters that may have caused misclassification on the basis of parameter saliency. It is also shown that fine-tuning the top ranking salient filters efficiently corrects misidentification on ImageNet. However, there is still a knowledge gap in terms of understanding why parameter saliency ranking can find the filters inducing misidentification. In this work, we attempt to bridge the gap by analyzing parameter saliency ranking from a statistical viewpoint, namely, extreme value theory. We first show that the existing work implicitly assumes that the gradient norm computed for each filter follows a normal distribution. Then, we clarify the relationship between parameter saliency and the score based on the peaks-over-threshold (POT) method, which is often used to model extreme values. Finally, we reformulate parameter saliency in terms of the POT method, where this reformulation is regarded as statistical anomaly detection and does not require the implicit assumptions of the existing parameter-saliency formulation. Our experimental results demonstrate that our reformulation can detect malicious filters as well. Furthermore, we show that the existing parameter saliency method exhibits a bias against the depth of layers in deep neural networks. In particular, this bias has the potential to inhibit the discovery of filters that cause misidentification in situations where domain shift occurs. In contrast, parameter saliency based on POT shows less of this bias.

Related papers

• Learning Probabilistic Filters with Strictly Proper Scoring Rules [8.61813579171868]
  We introduce the proper scoring ensemble filter (PSEF), an ensemble data assimilation method based on training an analysis map to approximate the filtering distribution.<n>We prove that, under a realizability assumption, the population objective is minimized by the true Bayesian filtering distribution.<n> Numerical experiments show that the learned filter accurately approximates challenging filtering distributions, including nonlinear, non-Gaussian, and multi-modal posteriors.
  arXiv  Detail & Related papers  (2026-06-25T01:04:21Z)
• Rational ANOVA Networks [12.085684184678348]
  Deep neural networks treat nonlinearities as fixed primitives, limiting both interpretability and control over the induced function class.<n>We propose the Rational-ANOVA Network (RAN), a foundational architecture grounded in functional ANOVA decomposition and Pad-style rational approximation.
  arXiv  Detail & Related papers  (2026-02-03T20:46:00Z)
• Why Machine Learning Models Systematically Underestimate Extreme Values II: How to Fix It with LatentNN [0.2700171473617699]
  Attenuation bias affects astronomical data-driven models.<n>We show that neural networks suffer from the same attenuation bias.<n>We introduce LatentNN, a method that jointly optimize network parameters and latent input values.
  arXiv  Detail & Related papers  (2025-12-29T01:59:10Z)
• Stability and performance guarantees for misspecified multivariate score-driven filters [0.0]
  We address the problem of tracking multivariate unobserved time-varying parameters under potential model misspecification. Specifically, we examine implicit and explicit score-driven (ISD and ESD) filters, which update parameter predictions using the gradient of the postulated logarithmic observation density. We highlight the superior stability and performance of ISD over ESD filters through extensive simulation studies.
  arXiv  Detail & Related papers  (2025-02-07T15:48:16Z)
• Graph Structure Learning with Interpretable Bayesian Neural Networks [10.957528713294874]
  We introduce novel iterations with independently interpretable parameters. These parameters influence characteristics of the estimated graph, such as edge sparsity. After unrolling these iterations, prior knowledge over such graph characteristics shape prior distributions. Fast execution and parameter efficiency allow for high-fidelity posterior approximation.
  arXiv  Detail & Related papers  (2024-06-20T23:27:41Z)
• Benign Overfitting in Deep Neural Networks under Lazy Training [72.28294823115502]
  We show that when the data distribution is well-separated, DNNs can achieve Bayes-optimal test error for classification. Our results indicate that interpolating with smoother functions leads to better generalization.
  arXiv  Detail & Related papers  (2023-05-30T19:37:44Z)
• Diffusion Causal Models for Counterfactual Estimation [18.438307666925425]
  We consider the task of counterfactual estimation from observational imaging data given a known causal structure. We propose Diff-SCM, a deep structural causal model that builds on recent advances of generative energy-based models. We find that Diff-SCM produces more realistic and minimal counterfactuals than baselines on MNIST data and can also be applied to ImageNet data.
  arXiv  Detail & Related papers  (2022-02-21T12:23:01Z)
• More is Less: Inducing Sparsity via Overparameterization [2.885175627590247]
  In deep learning it is common to over parameterize neural networks, that is, to use more parameters than training samples. Quite surprisingly, generalize the neural network via (stochastic) gradient descent leads to that very well. Our proof relies on analyzing a certain Bregman divergence of the flow.
  arXiv  Detail & Related papers  (2021-12-21T07:55:55Z)
• Batch Normalization Tells You Which Filter is Important [49.903610684578716]
  We propose a simple yet effective filter pruning method by evaluating the importance of each filter based on the BN parameters of pre-trained CNNs. The experimental results on CIFAR-10 and ImageNet demonstrate that the proposed method can achieve outstanding performance.
  arXiv  Detail & Related papers  (2021-12-02T12:04:59Z)
• Binary Classification of Gaussian Mixtures: Abundance of Support Vectors, Benign Overfitting and Regularization [39.35822033674126]
  We study binary linear classification under a generative Gaussian mixture model. We derive novel non-asymptotic bounds on the classification error of the latter. Our results extend to a noisy model with constant probability noise flips.
  arXiv  Detail & Related papers  (2020-11-18T07:59:55Z)
• Implicit Under-Parameterization Inhibits Data-Efficient Deep Reinforcement Learning [97.28695683236981]
  More gradient updates decrease the expressivity of the current value network. We demonstrate this phenomenon on Atari and Gym benchmarks, in both offline and online RL settings.
  arXiv  Detail & Related papers  (2020-10-27T17:55:16Z)
• Memorizing without overfitting: Bias, variance, and interpolation in over-parameterized models [0.0]
  The bias-variance trade-off is a central concept in supervised learning. Modern Deep Learning methods flout this dogma, achieving state-of-the-art performance.
  arXiv  Detail & Related papers  (2020-10-26T22:31:04Z)
• SCOP: Scientific Control for Reliable Neural Network Pruning [127.20073865874636]
  This paper proposes a reliable neural network pruning algorithm by setting up a scientific control. Redundant filters can be discovered in the adversarial process of different features. Our method can reduce 57.8% parameters and 60.2% FLOPs of ResNet-101 with only 0.01% top-1 accuracy loss on ImageNet.
  arXiv  Detail & Related papers  (2020-10-21T03:02:01Z)
• Dependency Aware Filter Pruning [74.69495455411987]
  Pruning a proportion of unimportant filters is an efficient way to mitigate the inference cost. Previous work prunes filters according to their weight norms or the corresponding batch-norm scaling factors. We propose a novel mechanism to dynamically control the sparsity-inducing regularization so as to achieve the desired sparsity.
  arXiv  Detail & Related papers  (2020-05-06T07:41:22Z)

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.