BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Bird sounds and machine learning - Dan Stowell (Queen Mary Univers
 ity of London)
DTSTART:20150305T111500Z
DTEND:20150305T121500Z
UID:TALK58326@talks.cam.ac.uk
CONTACT:Isak Herman
DESCRIPTION:Bird sound recordings contain a wealth of information - about 
 species\,\nindividuals\, and their interactions. Applying machine learning
  to such\ndata has applications in conservation\, animal behaviour researc
 h\, and\ncomparative linguistics. The rich structure of bird sounds is app
 arent\nto us as listeners\, but how could we "decode" it?\n\nI will descri
 be an approach to automatic classification of bird species\nwhich is desig
 ned to fit the characteristics of bird sounds: often\nnoisy\, often fast-m
 odulated\, often more than one bird\, and many\ncandidate species. This ap
 proach is currently being deployed in our\nWarblr bird recognition app.\n\
 nI will also describe recent work on more structured analysis of sound\nre
 cordings: tracking individuals in a sound scene\, and modelling bird\n"con
 versations".
LOCATION:Rainbow Seminar Room (SS03)\, Computer Laboratory
END:VEVENT
END:VCALENDAR