Professor Mykola Pechenizkiy visited NYU R/AI on October 22, 2024, and gave a talk on his recent work.

Benchmarking and Reasoning about Fairness in Machine Learning

Abstract: Machine learning (ML), including fairness-aware machine learning (fairML) is often formulated as an optimization problem–we search through the space of many possible models for some better ones. Benchmarking has been driving lots of progress in modern ML by facilitating efficient trial-and-error. Benchmarking is also one of the main means of measuring success in fairML literature. While many fairness metrics and fairML approaches that account for them have been proposed some of the basic questions remain open: Do we truly understand what fairML is optimizing for? Are there exceptions to the Goodhart’s Law? Does the end justify the means?

In his talk, Mykola reflected on these questions with a quest not only to develop better benchmarking of/for fairML, but also to put more efforts into the development of methods and tools for auditing fairML solutions, enabling different stakeholders to better understand and reason about how a fairML model generates predictions and where the potential unwanted biases come from.

Speaker’s bio: Mykola Pechenizkiy is a Full Professor at the department of Mathematics and Computer Science, Eindhoven University of Technology (TU/e), where he holds the Data Mining Chair. His core expertise and research interests are in predictive analytics and knowledge discovery from evolving data, and in their application to real-world problems in industry, medicine and education. At the Data Science Center Eindhoven, he leads the Customer Journey interdisciplinary research program aiming at developing techniques for informed and responsible analytics. Mykola Pechenizkiy received his PhD from the Computer Science and Information Systems department at the University of Jyväskylä, Finland in 2005. In addition to his work at TU/e, he is an Adjunct Professor in Data Mining for Industrial Applications at the Department of Mathematical Information Technology at the University of Jyväskylä.