Callisto: Entropy based test generation and data quality assessment for Machine Learning Systems

TL;DR

CALLISTO is a novel blackbox testing framework for ML systems that uses prediction uncertainty to generate test cases and assess data quality, significantly improving error detection and identifying mislabeled data.

Contribution

It introduces the first uncertainty-based blackbox test generation framework for ML classifiers, enhancing error detection and data quality assessment.

Findings

Increased error detection by up to 20 times using uncertainty.

Identified low quality and mislabeled data in real datasets.

Detected thousands of errors across four datasets.

Abstract

Machine Learning (ML) has seen massive progress in the last decade and as a result, there is a pressing need for validating ML-based systems. To this end, we propose, design and evaluate CALLISTO - a novel test generation and data quality assessment framework. To the best of our knowledge, CALLISTO is the first blackbox framework to leverage the uncertainty in the prediction and systematically generate new test cases for ML classifiers. Our evaluation of CALLISTO on four real world data sets reveals thousands of errors. We also show that leveraging the uncertainty in prediction can increase the number of erroneous test cases up to a factor of 20, as compared to when no such knowledge is used for testing. CALLISTO has the capability to detect low quality data in the datasets that may contain mislabelled data. We conduct and present an extensive user study to validate the results of CALLISTO on identifying low quality data from four state-of-the-art real world datasets.