Keynote Speaker

Ben Armstrong (University of Waterloo, Canada)

Title: Social Choice and Machine Learning: Past, Present, and Future

Abstract: The benefits of interdisciplinary research are becoming increasingly evident at the intersection of social choice and machine learning. Social choice offers tools for aggregating information from multiple ML models into a principled output, helps in pruning large models or ensembles of models, and may soon guide the fine-tuning of large language models. Conversely, machine learning provides theoretical frameworks for analyzing and interpreting social choice rules, developing new rules, and guiding analytical research. In this talk, I will explore the connections between social choice and machine learning over the past two centuries. Starting with a historical perspective, I will move through several current and ongoing projects, concluding with a discussion on how social choice can continue to enhance machine learning in the future.