Gerdus Benade (BU) will be speaking in-person on his paper:
Abstract:
Top-k recommendations are ubiquitous, but are they stable? We study whether, given complete information, buyers and sellers prefer to continue participating in a platform using top-k recommendations rather than pursuing off-platform transactions. When there are no constraints on the number of exposures, top-k recommendations are stable. However, stable k-recommendations may not exist when exposures are constrained, e.g., due to limited inventory or exposure opportunities. We show that maximizing total buyer welfare under unit exposure constraints is stable, Pareto optimal and swap-envy free in three restricted preference domains: orthogonal buyers, identical buyers, and buyers with dichotomous valuations. We generalize these results to arbitrary exposure constraints and formulate an integer program to find stable recommendations (when they exist). Experiments on three real-world datasets find that common recommendation strategies exhibit substantial instability and envy. Among them, maximizing expected buyer welfare leads to the most stable outcomes.
We will meet as usual, 1pm-2pm EST in SEC 1.413, and streamed via Zoom at the usual link:
https://harvard.zoom.us/j/95184948637?pwd=bXBIc2U5MEZ0QmRUb01WQ0o0SXRCdz09
LOCATION:in SEC 1.413, and streamed via Zoom here: https://harvard.zoom.us/j/95184948637?pwd=bXBIc2U5MEZ0QmRUb01WQ0o0SXRCdz09 STATUS:CONFIRMED DTSTART:20221007T170000Z DTEND:20221007T183000Z END:VEVENT END:VCALENDAR