Deep Learning for Audio Classification
Wenwu Wang ( University of Surrey )
Audio scene analysis, event detection and tagging have attracted increasing interest recently, with a variety of potential applications in security surveillance, and acoustic sensing for smart homes and cities. This talk will present some recent and new development for several challenges related to this topic, including data challenges (e.g. DCASE 2016 & 2017), acoustic modelling, feature learning, and dealing with weakly labelled data using deep learning. We will show some latest results including the results of our proposed algorithms and some benchmark methods. We will also use some sound demos to show the potential applications of these algorithms.
Wenwu Wang is a Reader in Signal Processing within the Centre for Vision Speech and Signal Processing, University of Surrey, Guildford, U.K.. His current research interests include blind signal processing, sparse signal processing, audio-visual signal processing, machine learning and perception, machine audition (listening), and statistical anomaly detection. He has (co)-authored over 200 publications in these areas. Dr. Wang is an Associate Editor for the IEEE TRANSACTIONS ON SIGNAL PROCESSING. He is a Member of the Ministry of Defence University Defence Research Collaboration in Signal Processing (since 2009), a Member of the BBC Audio Research Partnership (since 2011), and a Member of the BBC Data Science Research Partnership (since 2017). He was a Tutorial Speaker for ICASSP 2013 and UDRC Summer School 2014, 2015, 2016, 2017, and SpaRTan/MacSeNet Spring School 2016, and London Intelligent Sensing Summer School 2017.