Export Reviews, Discussions, Author Feedback and Meta-Reviews

Assigned Reviewer 1: *Quality* The paper is nicely written and well structured. Regarding the experimental results: by Theorem 4, it would be very interesting to see how the proposed RAC algorithm competes with existing algorithms *empirically* on the pool-based active learning problem (e.g., [3][5][7], or [6] under the bounded noise setting), given the fact that the version space reduction function and the Gibbs error function both satisfy MLRB condition, pointwise submodularity condition, and adaptive submodularity condition. Having these results will certainly make the paper stronger. A We thank reviewer for feedback. We intend to compare with algorithms on active learning problems in an extended version of this paper.