mlscorecheck: Testing the consistency of reported performance scores and experiments in machine learning
This addresses the problem of unreliable reported results for researchers and practitioners in machine learning, though it is incremental as it builds on existing validation methods.
The paper tackles the reproducibility crisis in AI by developing numerical techniques to identify inconsistencies between reported performance scores and experimental setups in machine learning, resulting in an open-source package called mlscorecheck that includes test bundles for detecting systematic flaws in fields like retina image processing.
Addressing the reproducibility crisis in artificial intelligence through the validation of reported experimental results is a challenging task. It necessitates either the reimplementation of techniques or a meticulous assessment of papers for deviations from the scientific method and best statistical practices. To facilitate the validation of reported results, we have developed numerical techniques capable of identifying inconsistencies between reported performance scores and various experimental setups in machine learning problems, including binary/multiclass classification and regression. These consistency tests are integrated into the open-source package mlscorecheck, which also provides specific test bundles designed to detect systematically recurring flaws in various fields, such as retina image processing and synthetic minority oversampling.