Summary of Provably Reliable Conformal Prediction Sets in the Presence Of Data Poisoning, by Yan Scholten et al.
Provably Reliable Conformal Prediction Sets in the Presence of Data Poisoning
by Yan Scholten, Stephan Günnemann
First submitted to arxiv on: 13 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
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 paper proposes reliable prediction sets (RPS) to mitigate the effects of data poisoning attacks on conformal prediction models. Conformal prediction provides model-agnostic and distribution-free uncertainty quantification through prediction sets that include the ground truth with any user-specified probability. However, existing methods are not reliable under poisoning attacks where adversaries manipulate both training and calibration data. RPS addresses this issue by introducing smoothed score functions to aggregate predictions of classifiers trained on distinct partitions of the training data, and constructing multiple prediction sets, each calibrated on distinct subsets of the calibration data. The approach mitigates the influence of datapoints in the training and calibration data on the final prediction set. Experimental results demonstrate strong reliability while maintaining utility and preserving coverage on clean data. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps make sure that computer programs can be trusted when they’re not 100% sure about what they’re saying. This is important because sometimes people try to trick these programs by giving them fake information. The authors come up with a way to fix this problem called reliable prediction sets (RPS). RPS makes sure that even if the program gets some bad information, it will still make good predictions most of the time. |
Keywords
» Artificial intelligence » Probability
