Summary of Longitudinal Counterfactuals: Constraints and Opportunities, by Alexander Asemota and Giles Hooker
Longitudinal Counterfactuals: Constraints and Opportunities
by Alexander Asemota, Giles Hooker
First submitted to arxiv on: 29 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computers and Society (cs.CY)
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 Counterfactual explanations are a crucial aspect of providing recourse to data subjects in machine learning models. However, current methods often produce counterfactuals that are impractical or impossible to achieve by the subject, rendering them less effective. The importance of plausibility when using counterfactuals for algorithmic recourse is widely acknowledged, but quantifying ground truth plausibility remains a significant challenge. This paper proposes leveraging longitudinal data to evaluate and enhance the plausibility of counterfactuals. A novel metric is developed to compare longitudinal changes with counterfactual differences, enabling the assessment of how similar the proposed counterfactual is to previous observed alterations. Furthermore, this metric is used to generate plausible counterfactuals that align better with real-world scenarios. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper talks about a way to make sure the fake explanations we give to people are believable and realistic. Right now, these “what if” stories can be pretty far-fetched and not very helpful for people who want to understand why they were denied a loan or didn’t get hired. The researchers came up with a new way to check how good these fake explanations are by comparing them to real changes that happened in the past. This helps make sure we’re giving people more realistic and useful “what if” stories. |
Keywords
* Artificial intelligence * Machine learning
