Summary of Look Around and Find Out: Ood Detection with Relative Angles, by Berker Demirel et al.
Look Around and Find Out: OOD Detection with Relative Angles
by Berker Demirel, Marco Fumero, Francesco Locatello
First submitted to arxiv on: 6 Oct 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 This paper proposes a novel angle-based metric for out-of-distribution (OOD) detection in deep learning systems. The method computes angles between feature representations and decision boundaries relative to the mean of in-distribution features, serving as an effective discriminative factor between ID and OOD data. The authors demonstrate state-of-the-art performance on CIFAR-10 and ImageNet benchmarks, achieving a reduction in FPR95 by 0.88% and 7.74%, respectively. The score function is compatible with existing feature space regularization techniques and has scale-invariance properties, enabling the creation of an ensemble of models for OOD detection via simple score summation. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps create better artificial intelligence systems that can avoid making mistakes when they encounter data they haven’t seen before. Usually, AI systems do well with the data they were trained on, but they struggle when faced with new or different types of data. The researchers developed a new way to detect when AI is looking at unfamiliar data and proposed an angle-based metric for this purpose. They tested their method on two big datasets, CIFAR-10 and ImageNet, and showed that it outperformed other methods in detecting unknown data. |
Keywords
» Artificial intelligence » Deep learning » Regularization
