Summary of On the Robustness Of Kernel Goodness-of-fit Tests, by Xing Liu et al.
On the Robustness of Kernel Goodness-of-Fit Tests
by Xing Liu, François-Xavier Briol
First submitted to arxiv on: 11 Aug 2024
Categories
- Main: Machine Learning (stat.ML)
- Secondary: Machine Learning (cs.LG); Statistics Theory (math.ST); Methodology (stat.ME)
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 paper addresses the limitation of traditional goodness-of-fit testing by developing a robust kernel goodness-of-fit test for probabilistic models. In machine learning, it is common to assume that all models are wrong, but this does not mean that the model is not good enough for a specific task. The authors show that existing kernel-based tests lack robustness, which is critical in today’s data-driven applications. They propose a novel approach using kernel Stein discrepancy balls, which incorporates perturbation models such as Huber contamination and density uncertainty bands. This work has implications for ensuring the quality of probabilistic models in various domains. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper explores ways to test whether data fits a model well enough, even when the model is not perfect. Currently, we reject most models because they’re not exactly right. But that doesn’t mean the model can’t be useful. The authors investigate why current tests for this problem aren’t working and develop a new approach to make sure our models are good enough. |
Keywords
» Artificial intelligence » Machine learning
