Fair ML Book Workshop

Brian Hedden of the ANU and Katie Creel of Northeastern University organized a workshop on Fairness and Machine Learning: Limitations and Opportunities on January 23rd at Stanford University. Fairness and Machine Learning is co-authored by Solon Barocas, Moritz Hardt, Arvind Narayanan. It is one of the first textbooks to focus exclusively on normative issues regarding the design and deployment of automated systems.

The workshop consisted of a series of presentations by philosophers that focused on individual book chapters. Henrik Kugelberg and Diana Acosta Navas discussed chapter two and the legitimacy of algorithmic decision making; Brian Hedden discussed chapter three and formal criteria for fair classification; Valerie Soon and Jamie Michelson discussed chapter four and Relative notions of fairness; Atoosa Kasirzadeh discussed chapter seven and structural injustice; Katie Creel discussed chapter eight and a broader view of discrimination; and Ramon Alvarado discussed chapter nine and issues relating to global inequality and datasets.

Solon and Arvind responded to the presenters and audience questions in person. The authors reported benefiting from the exchange and remarked that the workshop was a rare opportunity for interdisciplinary collaboration in that it provided them the opportunity to respond to exclusively philosophical criticism.

Seth Lazar and Rob Reich offered the general organizational direction for the workshop, with the ANU’s MINT Lab and the Stanford McCoy Family Center for Ethics in Society providing funding. A draft of Fairness and Machine Learning is available online at https://fairmlbook.org/. A hardcover version of the textbook has been published by the MIT Press and is scheduled for release in 2023.