Summary of Learning with Norm Constrained, Over-parameterized, Two-layer Neural Networks, by Fanghui Liu et al.
Learning with Norm Constrained, Over-parameterized, Two-layer Neural Networks
by Fanghui Liu, Leello Dadi, Volkan Cevher
First submitted to arxiv on: 29 Apr 2024
Categories
- Main: Machine Learning (stat.ML)
- 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 paper investigates suitable function spaces for over-parameterized neural networks, specifically addressing the curse of dimensionality when modeling ReLU neurons. By using the path norm or Barron norm as a measure of model complexity, the authors establish width-independence sample complexity bounds, enabling uniform convergence guarantees. The study also derives improved metric entropy results via convex hull techniques, demonstrating separation from kernel methods. These findings have implications for generalization properties and are relevant to applications in machine learning. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper looks at how neural networks can be used to model functions without running into a problem called the curse of dimensionality. They explore two types of norms that help measure how complex a network is, and show that these norms can help us understand when we need more data or more complexity in our models. This research has important implications for how well machine learning models generalize to new situations. |
Keywords
» Artificial intelligence » Generalization » Machine learning » Relu
