Summary of Adapting Conformal Prediction to Distribution Shifts Without Labels, by Kevin Kasa et al.
Adapting Conformal Prediction to Distribution Shifts Without Labels
by Kevin Kasa, Zhiyu Zhang, Heng Yang, Graham W. Taylor
First submitted to arxiv on: 3 Jun 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 paper proposes two new methods, ECP and EACP, to improve conformal prediction (CP) in classification tasks. CP provides prediction sets with guaranteed coverage rates, but often assumes exchangeable data, which is not realistic due to distribution shifts. The authors aim to address this challenge using only unlabeled test data. They adjust the score function in CP based on the base model’s uncertainty, leading to consistent improvements over existing baselines and nearly matching supervised algorithms’ performance. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps machine learning models predict sets with guaranteed coverage rates. Normally, these predictions assume that data is equally likely to come from any group, but this isn’t always true. The authors want to fix this problem using only the test data without labels. They change how CP works based on the base model’s uncertainty about the unlabeled test data. This makes their methods better than existing ones and almost as good as those that use labeled data. |
Keywords
» Artificial intelligence » Classification » Machine learning » Supervised
