Summary of Using Stratified Sampling to Improve Lime Image Explanations, by Muhammad Rashid et al.
Using Stratified Sampling to Improve LIME Image Explanations
by Muhammad Rashid, Elvio G. Amparore, Enrico Ferrari, Damiano Verda
First submitted to arxiv on: 26 Mar 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: None
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 This paper explores the application of a stratified sampling technique to LIME Image, a model-agnostic explainable AI method for computer vision tasks. The goal is to mitigate artifacts generated by typical Monte Carlo sampling, which can lead to inadequate explanations due to undersampling of the dependent variable. By leveraging Shapley theory’s insights on sample relevance and undersampling, this approach derives formulas for an unbiased stratified sampling estimator. Experimental results demonstrate the effectiveness of this method in improving explainability. Keywords: LIME Image, model-agnostic, computer vision, stratified sampling, Monte Carlo sampling, Shapley theory. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research paper is about making artificial intelligence (AI) more understandable. Right now, AI models are good at predicting things like what’s in a picture, but it’s hard to figure out why they made those predictions. The authors of this paper want to change that by using a new way to sample data that makes the explanations better. They’re building on some old ideas from economics and statistics to make sure their approach is fair and accurate. The results show that this method works well in computer vision tasks, like recognizing objects in pictures. |
