Summary of Adapted-moe: Mixture Of Experts with Test-time Adaption For Anomaly Detection, by Tianwu Lei et al.
Adapted-MoE: Mixture of Experts with Test-Time Adaption for Anomaly Detection
by Tianwu Lei, Silin Chen, Bohan Wang, Zhengkai Jiang, Ningmu Zou
First submitted to arxiv on: 9 Sep 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
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 an Adapted-MoE model for unsupervised anomaly detection that handles the variation in feature distribution for normal samples within the same category. Existing methods learn a single decision boundary, neglecting this variation and resulting in biased performance when applied to unseen test sets. The proposed model uses a routing network to route same-category samples into subclasses, followed by expert models learning separate representations and constructing independent decision boundaries. A test-time adaptation mechanism is also introduced to eliminate bias between the test sample representation and the feature distribution learned by the expert model. Experimental results on the Texture AD benchmark show significant improvements in performance compared to current state-of-the-art methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us find things that don’t belong in a group of normal things. It’s like looking for someone who doesn’t fit in with the crowd. Right now, most ways we do this only learn one way to tell if something is an outlier or not, and that’s based on how normal things are represented. But in real life, there can be many different normal things within the same category. So, this paper proposes a new way to do this by dividing up normal things into smaller groups, learning special ways to recognize each group, and then combining those ways to find outliers. It seems to work really well on a test dataset. |
Keywords
» Artificial intelligence » Anomaly detection » Unsupervised
