Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
This paper proposes a novel idea of multi-armed bandit formulation to tiered interviews. While selecting top K arms is not a new problem, application to interviewing and selecting set of candidates seems novel. The technical contributions and theory seem non-trivial, sound, and feasible. There are few aspects that need to be addressed as per reviewer comments: 1. Comparing the results of the proposed algorithm to a greedy baseline 2. Improving the clarity on making a connection between tiered interviewing and the proposed bandit formulation. 3. As suggested by reviewer 1, making the experimental setup more realistic and clarifying how to interpret the comparison of the proposed approach with the baselines (which are designed with different objectives) is critical. All in all, I think this is a borderline paper with some interesting ideas.