Summary of Enhancing Counterfactual Explanation Search with Diffusion Distance and Directional Coherence, by Marharyta Domnich et al.
Enhancing Counterfactual Explanation Search with Diffusion Distance and Directional Coherence
by Marharyta Domnich, Raul Vicente
First submitted to arxiv on: 19 Apr 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
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 tackles the challenge of generating human-centric explanations for AI models’ predictions by proposing a new approach to finding counterfactual explanations. Inspired by insights from human cognition, the authors develop a methodology that incorporates two novel biases: diffusion distance and directional coherence. The former emphasizes data connectivity and actionability in searching for feasible counterfactuals, while the latter enables the generation of explanations that align with marginal predictions based on feature changes. The proposed method, CoDiCE, is evaluated against existing methods on synthetic and real datasets with continuous and mixed-type features, demonstrating its effectiveness. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps AI models explain their predictions in a way that humans can understand by finding counterfactuals. It’s like trying to figure out why a model made a certain prediction by changing one thing at a time. The authors came up with two new ideas: “diffusion distance” and “directional coherence”. These ideas help the model find a path between what actually happened and what could have happened if something was different. They tested their idea, called CoDiCE, on some datasets and showed that it works better than other methods. |
Keywords
» Artificial intelligence » Diffusion
