Advances in Multi-label Classification

Abstract:
Multi-label classification has attracted much attention over the last few years, and its mention in the literature had expanded prolifically. In multi-label classification, each data instance may be associated with
multiple classes, as opposed to a single class label (as in the traditional classification task). This context arises naturally in a wide variety of domains (text categorisation, tagging images and other media, medical diagnosis, and bioinformatics, to name a few). The main challenge is to model dependencies between labels, which must be done efficiently for the learning algorithm to scale up to settings involving large datasets or data streams. This talk will review multi-label classification, particularly advances specific to the approach of 'classifier chains' which has attracted much recent interest; and multi-label learning in the context of data streams.

Bio:
Jesse Read is a lecturer (until recently postdoc) at the University Carlos III of Madrid, Spain. In 2010 he completed his PhD in the University of Waikato in New Zealand, with a thesis on scalable multi-label classification. Throughout his postdoc he has continued work on multi-label classification and related tasks such as multi-output prediction; but has also done research in the areas of data-stream classification and wireless sensor networks.