Adversarial Learning for Feature Shift Detection and Correction

TL;DR

This paper introduces an adversarial learning approach to detect and correct feature shifts in datasets, improving upon existing statistical and neural methods, with practical applications in multi-sensor and structured data.

Contribution

It presents a novel adversarial learning framework for localizing and correcting feature shifts, enhancing data integrity in various real-world datasets.

Findings

Outperforms existing statistical and neural methods in feature shift correction.

Effective with mainstream classifiers like random forests and gradient boosting.

Applicable to diverse data types including multi-sensor and biomedical data.

Abstract

Data shift is a phenomenon present in many real-world applications, and while there are multiple methods attempting to detect shifts, the task of localizing and correcting the features originating such shifts has not been studied in depth. Feature shifts can occur in many datasets, including in multi-sensor data, where some sensors are malfunctioning, or in tabular and structured data, including biomedical, financial, and survey data, where faulty standardization and data processing pipelines can lead to erroneous features. In this work, we explore using the principles of adversarial learning, where the information from several discriminators trained to distinguish between two distributions is used to both detect the corrupted features and fix them in order to remove the distribution shift between datasets. We show that mainstream supervised classifiers, such as random forest or gradient boosting trees, combined with simple iterative heuristics, can localize and correct feature shifts, outperforming current statistical and neural network-based techniques. The code is available at https://github.com/AI-sandbox/DataFix.