Summary of Class Relevance Learning For Out-of-distribution Detection, by Butian Xiong et al.
Class Relevance Learning For Out-of-distribution Detection
by Butian Xiong, Liguang Zhou, Tin Lun Lam, Yangsheng Xu
First submitted to arxiv on: 21 Sep 2023
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
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 tackles the challenge of out-of-distribution (OOD) detection in image classification models. Currently, existing methods like max logits are insufficient as they disregard the relationships between classes. The proposed class relevance learning method addresses this issue by leveraging interclass relationships within the OOD pipeline. This approach significantly improves OOD detection capabilities compared to state-of-the-art alternatives. The paper’s results demonstrate the effectiveness of the method on various datasets, including Near OOD and Far OOD datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research helps us make better image classification models that can work well in real-world situations. Right now, these models have trouble identifying new classes they weren’t trained to recognize. To fix this, scientists are working on a way to detect when an image is from a class the model hasn’t seen before. The usual approach doesn’t take into account how different classes relate to each other. This new method fixes that by using relationships between classes to improve detection. It works really well and can be applied to many different datasets. |
Keywords
* Artificial intelligence * Image classification * Logits
