Summary of Rethinking Data Shapley For Data Selection Tasks: Misleads and Merits, by Jiachen T. Wang et al.
Rethinking Data Shapley for Data Selection Tasks: Misleads and Merits
by Jiachen T. Wang, Tianji Yang, James Zou, Yongchan Kwon, Ruoxi Jia
First submitted to arxiv on: 6 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: 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 proposed study delves into the inconsistent performance of Data Shapley in data-centric machine learning (ML) research, particularly in data selection applications. By introducing a hypothesis testing framework, researchers uncover that without specific constraints on utility functions, Data Shapley’s performance can be no better than random selection. However, they also identify a class of utility functions where Data Shapley optimally selects data, leading to the development of a heuristic for predicting its effectiveness in data selection tasks. Experimental results validate these findings, providing valuable insights into when Data Shapley may succeed or fail. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study looks at how well Data Shapley works in choosing the right data for machine learning projects. Researchers found that without special rules, Data Shapley does about as well as picking random data. However, they discovered a type of utility function where Data Shapley really shines, and created a simple rule to figure out when it will work best. This helps us understand when we can trust Data Shapley or not. |
Keywords
» Artificial intelligence » Machine learning
