Reviews: Active Learning with Oracle Epiphany

Overall I enjoyed reading this paper. It is very well written, and the algorithms and analysis seem to be natural modifications of these existing approaches to active learning. The theoretical issues that arise in handling these abstentions and quantifying their effect on the query complexity are at times nontrivial, and are handled in elegant and appropriate ways. I suspect that not many people are aware of this problem, but it is quite well motivated in the paper, and seems to be a good problem to study. The specific theoretical model proposed for this phenomenon is, however, a bit toy-like.