Summary of Logit Scaling For Out-of-distribution Detection, by Andrija Djurisic et al.
Logit Scaling for Out-of-Distribution Detection
by Andrija Djurisic, Rosanne Liu, Mladen Nikolic
First submitted to arxiv on: 2 Sep 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); 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 presents a novel approach to out-of-distribution (OOD) data detection in machine learning models. The proposed method, Logit Scaling (LTS), does not require access to the training data distribution and can be applied to any pre-trained network without modifying it. LTS scales the logits to effectively distinguish between in-distribution (ID) and OOD samples. The authors evaluate their method on a range of benchmarks across various scales, including CIFAR-10, CIFAR-100, ImageNet, and OpenOOD, covering 3 ID datasets, 14 OOD datasets, and 9 model architectures. The results demonstrate state-of-the-art performance, robustness, and adaptability across different architectures. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps machines learn better by finding out-of-distribution data. Right now, methods for doing this often require a lot of extra work or access to information that might not be available. This new method, called Logit Scaling (LTS), can be used on any machine learning model without changing it. LTS makes it easy to tell the difference between normal data and weird data. The authors tested their method on many different datasets and models, and it worked really well. |
Keywords
» Artificial intelligence » Logits » Machine learning
