Summary of Interpretable Network Visualizations: a Human-in-the-loop Approach For Post-hoc Explainability Of Cnn-based Image Classification, by Matteo Bianchi et al.
Interpretable Network Visualizations: A Human-in-the-Loop Approach for Post-hoc Explainability of CNN-based Image Classification
by Matteo Bianchi, Antonio De Santis, Andrea Tocchetti, Marco Brambilla
First submitted to arxiv on: 6 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computer Vision and Pattern Recognition (cs.CV)
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 Medium Difficulty summary: This paper addresses the need for transparent and explainable image classification models by introducing a post-hoc method that breaks down the feature extraction process of Convolutional Neural Networks (CNNs). The approach generates layer-wise feature representations as saliency maps, which are weighted to show the importance of each feature. To further enhance explanations, the authors collect textual labels through gamified crowdsourcing and process them using NLP techniques and Sentence-BERT. Additionally, the paper proposes a method for generating global explanations by aggregating labels across multiple images. This research has implications for detecting biases and errors in image classification models and establishing trust with users. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Low Difficulty summary: Imagine if you could understand how a computer program makes decisions about what it sees in pictures. This is important because we want to make sure the programs don’t make mistakes or show bias. The authors of this paper came up with a new way to do just that by breaking down how a special kind of computer program, called a Convolutional Neural Network (CNN), works when it looks at an image. They also figured out ways to explain what the program is looking for in the picture and why it made certain decisions. This research could help us create more trustworthy programs that don’t make mistakes or show bias. |
Keywords
» Artificial intelligence » Bert » Cnn » Feature extraction » Image classification » Neural network » Nlp
