Summary of An Axiomatic Approach to Model-agnostic Concept Explanations, by Zhili Feng et al.
An Axiomatic Approach to Model-Agnostic Concept Explanations
by Zhili Feng, Michal Moshkovitz, Dotan Di Castro, J. Zico Kolter
First submitted to arxiv on: 12 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
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 proposes a novel approach to concept explanations that can be applied to any machine learning model, rather than being tailored to specific models. The proposed method satisfies three natural axioms: linearity, recursivity, and similarity, making it a versatile tool for understanding how human-interpretable concepts impact model predictions. By connecting the new method with previous approaches, this paper provides insight into their varying semantic meanings. Experimentally, the authors demonstrate the utility of the new method in different scenarios, including model selection, optimizer selection, and model improvement using prompt editing for zero-shot vision language models. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us understand how to explain why a machine learning model makes certain predictions. Right now, most methods are designed for specific types of models, but this new approach works with any type of model. The authors came up with three rules (linearity, recursivity, and similarity) that their method follows. They also show how their method relates to other ways people try to explain why a model makes certain predictions. To test it out, the authors used their method in different situations to see if it’s useful for things like choosing the best model or improving an existing one. |
Keywords
* Artificial intelligence * Machine learning * Prompt * Zero shot
