Neural Information Processing Systems http://nips.cc/

Causal inference in randomized experiments typically assumes that the units of randomization and the units of analysis are one and the same. In some applications, however, these two roles are played by distinct entities linked by a bipartite graph. The key challenge in such bipartite settings is how to avoid interference bias, which would typically arise if we simply randomized the treatment at the level of analysis units. One effective way of minimizing interference bias in standard experiments is through cluster randomization, but this design has not been studied in the bipartite setting where conventional clustering schemes can lead to poorly powered experiments. This paper introduces a novel clustering objective and a corresponding algorithm that partitions a bipartite graph so as to maximize the statistical power of a bipartite experiment on that graph. Whereas previous work relied on balanced partitioning, our formulation suggests the use of a correlation clustering objective. We use a publicly-available graph of Amazon user-item reviews to validate our solution and illustrate how it substantially increases the statistical power in bipartite experiments.

We would like to thank all three reviewers for their careful review and constructive feedback. " In the introduction the connection to optimal experimental design is alluded to. " Thank you for this feedback. "It is not clear what the quality of the approximation is for the heuristic presented [...]" We agree that the quality of the "What is the space complexity of the proposed heuristic?" "Why are other correlation clustering heuristics not compared against in the experiment section?" "There are no results given for the variance of the estimators / designs [...]" Thank you for this feedback.