Output details

Combining Long Short-Term Memory and Dynamic Bayesian Networks for Incremental Emotion-Sensitive Artificial Listening

<24>Introduces for the first time the combination of a long-short term memory neural architecture as developed at TUM in tandem operation with dynamic Bayesian networks for combined speech and emotion recognition. It triggered a wave of manifold papers in a similar direction. With the help of the methods, the authors later won the CHiME 2013 Challenge (http://spandh.dcs.shef.ac.uk/chime_challenge/) held in conjunction with the flagship ICASSP conference, and the MIREX 2010 onset and tempo challenges (http://www.music-ir.org/mirex/wiki/2010:MIREX2010_Results) held in conjunction with the flagship ISMIR conference. The technique also helped win the IEEE/Humaine ACII 2009 best technical demo award in 2009.