Math/Stats Colloquium: Anna Haensch, Tufts University
Tue, May 2, 2023 • 4:00pm - 5:00pm (1h) • CMC 206
Title: OpArt: An Equity Based Recommender System for Optimal Art Curation
Abstract: The placement of art in public spaces can have a significant impact on who feels a sense of belonging in a space. In cities, public art communicates whose interests and culture are being favored. On college campuses, which act like miniature cities, the same is true. In this joint work with Abiy Tasissa and Dina Deitsch, we explore a curatorial tool for selecting public art in a way that supports inclusive spaces. Drawing on Schelling’s model of segregation, we use an assignment derived from the graph matching algorithm and optimized using projected gradient descent to propose an allocation of artwork in a way that de-prioritizes “in-group” preferences, by satisfying minimum representation and exposure criteria. We attempt to bring notions of algorithmic fairness and equity based metrics into this allocation problem and we will discuss the potential pitfalls of this approach from both a curatorial and equity standpoint.