Summary of Learning Recourse Costs From Pairwise Feature Comparisons, by Kaivalya Rawal and Himabindu Lakkaraju
Learning Recourse Costs from Pairwise Feature Comparisons
by Kaivalya Rawal, Himabindu Lakkaraju
First submitted to arxiv on: 20 Sep 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Machine Learning (stat.ML)
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 paper proposes a novel technique for incorporating user input when learning and inferring user preferences in black-box machine learning models. The goal is to provide users with actionable recourse by modifying individual features based on their personal preferences. The traditional approach requires an exhaustive set of tuples associating each feature to its cost of modification, which is challenging to obtain through direct human surveying. Instead, the authors propose using the Bradley-Terry model to automatically infer feature-wise costs using non-exhaustive human comparison surveys. Users provide inputs comparing entire recourses with all candidate feature modifications, determining which recourses are easier to implement relative to others without explicit quantification of their costs. The paper demonstrates efficient learning of individual feature costs using MAP estimates and shows that non-exhaustive human surveys can learn an exhaustive set of feature costs, where each feature is associated with a modification cost. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper helps us understand how people’s preferences are used to make machine learning models better. It’s like trying to find the best way to modify features in a model based on what someone wants. Usually, we need lots of information about how hard it is to change each feature. But that’s hard to get from people directly. So, this paper shows us how to use something called the Bradley-Terry model to figure out how hard it is to change each feature by asking people to compare different ways to modify a model. |
Keywords
* Artificial intelligence * Machine learning
