Summary of Navigate Beyond Shortcuts: Debiased Learning Through the Lens Of Neural Collapse, by Yining Wang et al.
Navigate Beyond Shortcuts: Debiased Learning through the Lens of Neural Collapse
by Yining Wang, Junjie Sun, Chenyue Wang, Mi Zhang, Min Yang
First submitted to arxiv on: 9 May 2024
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 The proposed framework avoids shortcut learning in biased classification tasks by inducing a Neural Collapse structure through prime training, leading to improved generalization performance. The approach leverages the symmetric and stable feature space established during neural network training, encouraging models to focus on intrinsic correlations rather than simple shortcuts. Experimental results demonstrate the effectiveness of this method on both synthetic and real-world datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper studies how neural networks work when given biased data. It shows that these networks often get stuck in a bad habit called “shortcut learning” which makes them poor at generalizing to new, unseen data. To fix this, the authors suggest a new way of training the models, using something called “prime training”. This helps the models focus on what’s really important and not just find easy ways out. |
Keywords
» Artificial intelligence » Classification » Generalization » Neural network
