Summary of Classification Metrics For Image Explanations: Towards Building Reliable Xai-evaluations, by Benjamin Fresz et al.
Classification Metrics for Image Explanations: Towards Building Reliable XAI-Evaluations
by Benjamin Fresz, Lena Lörcher, Marco Huber
First submitted to arxiv on: 7 Jun 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary In this paper, the authors tackle the issue of evaluating saliency methods in computer vision, specifically deep neural networks. Saliency methods provide feature attribution scores for input images, but their evaluation remains a problem due to the lack of ground truth. To address this, new metrics are developed and common saliency methods are benchmarked on ImageNet. The authors also propose a scheme for evaluating the reliability of these metrics based on psychometric testing concepts. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about making computer vision models more understandable. Right now, we can’t figure out why deep neural networks make certain decisions. To help with this, many methods have been created to explain what’s going on inside these models. One type of method is called saliency, which shows how different parts of an image are related to the model’s decision. The problem is that it’s hard to tell if these explanations are any good because we don’t know what the “right” explanation should look like. To solve this, the authors came up with new ways to measure how well saliency methods work and tested some popular methods on a big image dataset. |
