Summary of On Model Extrapolation in Marginal Shapley Values, by Ilya Rozenfeld
On Model Extrapolation in Marginal Shapley Values
by Ilya Rozenfeld
First submitted to arxiv on: 17 Dec 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 addresses the issue of reliable explainability methods for complex machine learning models, particularly those based on Shapley values. It highlights the limitations of two commonly used approaches: conditional and marginal. The authors show that the conditional approach is flawed due to implicit assumptions of causality, while the marginal approach can lead to model extrapolation. They propose a new method that avoids model extrapolation by using marginal averaging and incorporating causal information, which replicates causal Shapley values. The paper demonstrates this approach on a real-world data example. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper is about making complex machine learning models more understandable. It talks about two ways to do this called Shapley values. One way has problems because it assumes things that aren’t true. The other way can also cause issues when the model tries to predict something new. The authors suggest a new method that fixes these problems and works better. They test their idea on real data. |
Keywords
» Artificial intelligence » Machine learning
