Summary of Hard to Explain: on the Computational Hardness Of In-distribution Model Interpretation, by Guy Amir et al.
Hard to Explain: On the Computational Hardness of In-Distribution Model Interpretation
by Guy Amir, Shahaf Bassan, Guy Katz
First submitted to arxiv on: 7 Aug 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computational Complexity (cs.CC); Logic in Computer Science (cs.LO)
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 investigates the interpretability of Machine Learning (ML) models, focusing on three key factors that affect their explainability: the type of model, form of explanation, and underlying distribution. The authors argue that considering the underlying distribution is crucial in avoiding biased and unhelpful explanations. They demonstrate the influence of this factor on interpretation complexity through two case studies: prediction models paired with an external out-of-distribution (OOD) detector, and those designed to generate socially aligned explanations. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper looks at how well machine learning models can be understood. It’s important because we need to know why these models make certain decisions. The authors found that three things affect how easy it is to understand an ML model: what type of model it is, how you explain its decisions, and the kind of data it’s trained on. They tested this by looking at two different types of models: ones that work with external detectors and ones designed to be socially fair. |
Keywords
» Artificial intelligence » Machine learning
