Reliable Evaluation of Attribution Maps in CNNs: A Perturbation-Based Approach
- URL: http://arxiv.org/abs/2411.14946v1
- Date: Fri, 22 Nov 2024 13:57:56 GMT
- Title: Reliable Evaluation of Attribution Maps in CNNs: A Perturbation-Based Approach
- Authors: Lars Nieradzik, Henrike Stephani, Janis Keuper,
- Abstract summary: We present an approach for evaluating attribution maps, which play a central role in interpreting predictions of convolutional neural networks (CNNGrads)
We show that the widely used insertion/deletion metrics are susceptible to distribution shifts that affect the reliability of the ranking.
Our method proposes to replace pixel modifications with adversarial perturbations, which provides a more robust evaluation framework.
- Score: 7.1606014219358425
- License:
- Abstract: In this paper, we present an approach for evaluating attribution maps, which play a central role in interpreting the predictions of convolutional neural networks (CNNs). We show that the widely used insertion/deletion metrics are susceptible to distribution shifts that affect the reliability of the ranking. Our method proposes to replace pixel modifications with adversarial perturbations, which provides a more robust evaluation framework. By using smoothness and monotonicity measures, we illustrate the effectiveness of our approach in correcting distribution shifts. In addition, we conduct the most comprehensive quantitative and qualitative assessment of attribution maps to date. Introducing baseline attribution maps as sanity checks, we find that our metric is the only contender to pass all checks. Using Kendall's $\tau$ rank correlation coefficient, we show the increased consistency of our metric across 15 dataset-architecture combinations. Of the 16 attribution maps tested, our results clearly show SmoothGrad to be the best map currently available. This research makes an important contribution to the development of attribution maps by providing a reliable and consistent evaluation framework. To ensure reproducibility, we will provide the code along with our results.
Related papers
- Less is More: Fewer Interpretable Region via Submodular Subset Selection [54.07758302264416]
This paper re-models the above image attribution problem as a submodular subset selection problem.
We construct a novel submodular function to discover more accurate small interpretation regions.
For correctly predicted samples, the proposed method improves the Deletion and Insertion scores with an average of 4.9% and 2.5% gain relative to HSIC-Attribution.
arXiv Detail & Related papers (2024-02-14T13:30:02Z) - Nearest Neighbor Guidance for Out-of-Distribution Detection [18.851275688720108]
We propose Nearest Neighbor Guidance (NNGuide) for detecting out-of-distribution (OOD) samples.
NNGuide reduces the overconfidence of OOD samples while preserving the fine-grained capability of the classifier-based score.
Our results demonstrate that NNGuide provides a significant performance improvement on the base detection scores.
arXiv Detail & Related papers (2023-09-26T12:40:35Z) - Unsupervised Landmark Discovery Using Consistency Guided Bottleneck [63.624186864522315]
We introduce a consistency-guided bottleneck in an image reconstruction-based pipeline.
We propose obtaining pseudo-supervision via forming landmark correspondence across images.
The consistency then modulates the uncertainty of the discovered landmarks in the generation of adaptive heatmaps.
arXiv Detail & Related papers (2023-09-19T10:57:53Z) - Okapi: Generalising Better by Making Statistical Matches Match [7.392460712829188]
Okapi is a simple, efficient, and general method for robust semi-supervised learning based on online statistical matching.
Our method uses a nearest-neighbours-based matching procedure to generate cross-domain views for a consistency loss.
We show that it is in fact possible to leverage additional unlabelled data to improve upon empirical risk minimisation.
arXiv Detail & Related papers (2022-11-07T12:41:17Z) - A Closer Look at Debiased Temporal Sentence Grounding in Videos:
Dataset, Metric, and Approach [53.727460222955266]
Temporal Sentence Grounding in Videos (TSGV) aims to ground a natural language sentence in an untrimmed video.
Recent studies have found that current benchmark datasets may have obvious moment annotation biases.
We introduce a new evaluation metric "dR@n,IoU@m" that discounts the basic recall scores to alleviate the inflating evaluation caused by biased datasets.
arXiv Detail & Related papers (2022-03-10T08:58:18Z) - Reliable Inlier Evaluation for Unsupervised Point Cloud Registration [26.883254695961682]
We propose a neighborhood consensus based reliable inlier evaluation method for robust unsupervised point cloud registration.
It is expected to capture the discriminative geometric difference between the source neighborhood and the corresponding pseudo target neighborhood for effective inlier distinction.
Under the unsupervised setting, we exploit the Huber function based global alignment loss, the local neighborhood consensus loss, and spatial consistency loss for model optimization.
arXiv Detail & Related papers (2022-02-23T03:46:42Z) - Evaluating Feature Attribution Methods in the Image Domain [7.852862161478641]
We investigate existing metrics and propose new variants of metrics for the evaluation of attribution maps.
We find that different attribution metrics seem to measure different underlying concepts of attribution maps.
We propose a general benchmarking approach to identify the ideal feature attribution method for a given use case.
arXiv Detail & Related papers (2022-02-22T15:14:33Z) - Accuracy on the Line: On the Strong Correlation Between
Out-of-Distribution and In-Distribution Generalization [89.73665256847858]
We show that out-of-distribution performance is strongly correlated with in-distribution performance for a wide range of models and distribution shifts.
Specifically, we demonstrate strong correlations between in-distribution and out-of-distribution performance on variants of CIFAR-10 & ImageNet.
We also investigate cases where the correlation is weaker, for instance some synthetic distribution shifts from CIFAR-10-C and the tissue classification dataset Camelyon17-WILDS.
arXiv Detail & Related papers (2021-07-09T19:48:23Z) - Normalized Label Distribution: Towards Learning Calibrated, Adaptable
and Efficient Activation Maps [0.0]
Vulnerability of models to data aberrations and adversarial attacks influences their ability to demarcate distinct class boundaries efficiently.
We study the significance of ground-truth distribution changes on the performance and generalizability of various state-of-the-art networks.
arXiv Detail & Related papers (2020-12-12T17:54:01Z) - Making Affine Correspondences Work in Camera Geometry Computation [62.7633180470428]
Local features provide region-to-region rather than point-to-point correspondences.
We propose guidelines for effective use of region-to-region matches in the course of a full model estimation pipeline.
Experiments show that affine solvers can achieve accuracy comparable to point-based solvers at faster run-times.
arXiv Detail & Related papers (2020-07-20T12:07:48Z) - Rethinking Localization Map: Towards Accurate Object Perception with
Self-Enhancement Maps [78.2581910688094]
This work introduces a novel self-enhancement method to harvest accurate object localization maps and object boundaries with only category labels as supervision.
In particular, the proposed Self-Enhancement Maps achieve the state-of-the-art localization accuracy of 54.88% on ILSVRC.
arXiv Detail & Related papers (2020-06-09T12:35:55Z)
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.