NeurIPS 2020

Online MAP Inference of Determinantal Point Processes


Meta Review

The paper proposes an algorithm for Online-DPPs, which has interesting approximation guarantees. While the work is an extension of Mahabadi et al, 2020, the required extensions are non-trivial. The paper shows improvements over a greedy baseline and is theoretically well-founded. Weaknesses of the paper include the experimental evaluation, where improvement over other methods was present but not very clear. The recommendation for NeurIPS is accept.