I used about 30 minutes of lecture time going through some state-of-the-art research outcomes presented at ACM Multimedia 2016 conference. Students seemed to enjoy it and some even tried to download the paper for more details of the research.

A keyword of this year’s ACM MM is deep learning and there have been some interesting discussions on whether we are truly learning anything through deep learning or merely engineering some increasingly complex CNNs to get some jobs done…

Courtesy to the rightful researchers.