Summary of Reassessing Evaluation Functions in Algorithmic Recourse: An Empirical Study From a Human-centered Perspective, by Tomu Tominaga et al.
Reassessing Evaluation Functions in Algorithmic Recourse: An Empirical Study from a Human-Centered Perspective
by Tomu Tominaga, Naomi Yamashita, Takeshi Kurashima
First submitted to arxiv on: 23 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
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 study investigates the foundational premise of algorithmic recourse, which aims to help individuals reverse adverse AI decisions by generating counterfactual action plans. The researchers examine whether minimizing a metric called “distance function” actually prompts people to accept and act upon suggested recourses. They conducted a user study with 362 participants and found that people’s acceptance of recourses didn’t correlate with the distance, while their willingness to take action peaked at the minimal distance but remained constant otherwise. This challenges the prevailing assumption in algorithmic recourse research and suggests rethinking evaluation functions for human-centered recourse generation. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper looks into how AI can help people fix bad decisions made by computers. It’s like a “undo” button, where AI finds ways to make things better. The researchers asked 362 people about this idea and found that it doesn’t quite work as expected. People might not even want to try the fixes if they’re too similar to what happened before. This makes us rethink how we measure whether these fixes are working. |
