BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Demodulation and time-frequency analysis as inference - Dr Rich Tu
 rner\, CUED.
DTSTART:20130530T130000Z
DTEND:20130530T140000Z
UID:TALK45602@talks.cam.ac.uk
CONTACT:Prof. Ramji Venkataramanan
DESCRIPTION:In this talk I will present a theoretical framework that links
  a set of widely used methods from signal processing to statistical infere
 nce procedures. This result will then be used as a conceptual springboard 
 \nto improve upon the classical methods.\n\nI will begin by describing a f
 amily of related inference problems that have optimal solutions correspond
 ing to the short-time Fourier transform (STFT)\, spectrogram\, filter bank
 \, and wavelet \nrepresentations of signals. The framework allows us to us
 e modern techniques from statistical inference to improve upon these class
 ical signal processing methods.\n\nI will show two examples where such an 
 approach has borne fruit. In the first example we use an inferential \next
 ension of the Hilbert method to produce high-quality approaches to joint a
 mplitude and frequency modulation of signals. The new approach is uncertai
 nty-aware and therefore noise robust which results in fewer \nartifacts. I
 n the second example we extend the STFT into an adaptive time-frequency an
 alysis using a hierarchical probabilistic model. The parameters of the new
  representation\, including the channel \ncentre-frequencies and bandwidth
 s\, can be learned directly from the signal. The adaptive representation c
 an be used to remove noise from signals and to impute missing data. Surpri
 singly\, the method is an \nexcellent model for naturally occurring audio 
 textures such as howling wind\, falling rain\, and running water.\n\nI wil
 l wrap up by discussing how we might bring the fields of signal processing
  and statistical inference closer together and the benefits and challenges
  of such a research effort.\n
LOCATION:LR11\, Engineering\, Department of
END:VEVENT
END:VCALENDAR