Summary of Efficient and Accurate Explanation Estimation with Distribution Compression, by Hubert Baniecki et al.
Efficient and Accurate Explanation Estimation with Distribution Compression
by Hubert Baniecki, Giuseppe Casalicchio, Bernd Bischl, Przemyslaw Biecek
First submitted to arxiv on: 26 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Machine Learning (stat.ML)
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 uncovers a theoretical connection between explanation estimation and distribution compression, enabling more accurate feature attributions, importance, and effects. The authors identify the limitations of standard i.i.d. sampling used in post-hoc explanation algorithms, which increase computational costs with growing data and model sizes. To address this issue, they introduce Compress Then Explain (CTE), a novel paradigm for sample-efficient explainability. CTE uses kernel thinning to compress distributions, reducing the number of samples needed to achieve accurate explanations while maintaining negligible computational overhead. This approach can be integrated into various explanation methods, providing a 2-3x speedup and comparable approximation error. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us understand how we can make machine learning models more transparent by improving the way they explain their decisions. Right now, it takes a lot of computation to figure out why a model made a certain decision. The authors found that this process gets slower as the data and model get bigger. They came up with a new way to do explanation estimation called Compress Then Explain (CTE), which is faster and more accurate. CTE works by reducing the amount of data needed for explanations, making it easier and faster to understand why models make certain decisions. |
Keywords
* Artificial intelligence * Machine learning
