Robust Learning from Untrusted Sources

TL;DR

This paper introduces a robust learning method that effectively handles untrusted, distributed, or private data sources, improving model reliability and accuracy in challenging data collection scenarios.

Contribution

The authors propose a novel statistical learning framework that automatically suppresses irrelevant or corrupted data from multiple sources, enhancing robustness in machine learning.

Findings

Significant performance improvements over existing robust methods.

Effective suppression of corrupted or irrelevant data.

Applicable to distributed and privacy-preserving data scenarios.

Abstract

Modern machine learning methods often require more data for training than a single expert can provide. Therefore, it has become a standard procedure to collect data from external sources, e.g. via crowdsourcing. Unfortunately, the quality of these sources is not always guaranteed. As additional complications, the data might be stored in a distributed way, or might even have to remain private. In this work, we address the question of how to learn robustly in such scenarios. Studying the problem through the lens of statistical learning theory, we derive a procedure that allows for learning from all available sources, yet automatically suppresses irrelevant or corrupted data. We show by extensive experiments that our method provides significant improvements over alternative approaches from robust statistics and distributed optimization.