State of the art methods for multi-label classification

Occasionally I am asked to review papers where a portion of the novelty is well outside my general area of expertise. Over the past couple months I've gotten one or two papers on multi-label classification, and I feel like I just don't know enough about it. I understand the baseline methods: one-vs-all, for transforming the multi-label problem to a series of binary problems and power-set, which treats the multi-label problem as a multi-class problem where there are 2^n possible classes of n labels. Beyond that, I'm a little lost. Any literature pointers would be much appreciated.