LSAS2008: Short Report about the Avantgarde in MIR
Tuesday, June 24th, 2008
On Saturday I attended the LSAS2008 workshop on Learning Semantics of Audio Signals. The workshop was initiated in 2006 by the guys of University of Magdeburg since they had been really curious to apply their deep machine learning know-how to new domains. Already the LSAS2006 was indeed a very interesting event since most of the papers have been more about the risky edge in MIR research. Now 2 years later the few but highly interesting papers were even more terrific… at least for me. Why? If watching the Robbie Williams DVD Live at Knebworth gives you hefty chills in certain passages around the more romantic songs such as Angels or Strong then you might think about building MIR tools to explain this. Such emotional recommenders taking into account full-fledge context seem to be the next hot thing, at least I had the impression at LSAS2008 following different presentations (including MIR dinosaur F.Pachet and the young and willing Ph.D candidates of the current decade).
So much for now, please contribute and spread the message that you can access and contribute chill-inducing songs at OPENEER. I should post this to the MIR mailinglist …