Summary of What Is the Right Notion Of Distance Between Predict-then-optimize Tasks?, by Paula Rodriguez-diaz et al.
What is the Right Notion of Distance between Predict-then-Optimize Tasks?
by Paula Rodriguez-Diaz, Lingkai Kong, Kai Wang, David Alvarez-Melis, Milind Tambe
First submitted to arxiv on: 11 Sep 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 research paper proposes a new approach to comparing datasets for machine learning models, particularly in Predict-then-Optimize (PtO) frameworks where decision regret minimization is used instead of prediction error minimization. The authors show that traditional dataset distances are insufficient in this context and introduce a novel distance metric that incorporates the impacts of downstream decisions. They demonstrate the effectiveness of their proposed distance measure in capturing adaptation success and predicting transferability across various PtO tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary In simple terms, this paper is about finding a better way to compare datasets for machine learning models. Traditionally, we use methods like prediction error minimization to evaluate how similar two datasets are. However, when we’re using predictions as inputs for other optimization tasks, we need a different approach. The authors propose a new method that takes into account the impact of these downstream decisions and show it works better than traditional methods. |
Keywords
» Artificial intelligence » Machine learning » Optimization » Transferability
