BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Perceiving is Believing: Bayesian inference in unexpected places. 
 - Maneesh Sahani\, Gatsby\, UCL
DTSTART:20120905T093000Z
DTEND:20120905T103000Z
UID:TALK39491@talks.cam.ac.uk
CONTACT:Dr. Cristina Savin
DESCRIPTION:The past century or more of research into perception has been\
 ndominated by two different computational metaphors: the Hemholtzian\ndesc
 ription of a process of inference about the external causes of\nsensations
  (in its modern guise often expressed in terms of Bayesian\nprobabilistic 
 reasoning)\, and a more mechanistic\, processing-based\nview often rooted 
 in engineering operations such as accumulation and\ncross-correlation.  Ma
 ny mid- and high-level perceptual phenomena seem\nto be well-understood in
 ferentially\, but the processing-based\nmetaphors have remained dominant f
 or perception based on 'early'\nsensory computations.\n\nHere I will prese
 nt evidence that inference plays a substantial role\nin two perceptual dom
 ains where most current theories rely on\nsignal-processing-based models. 
  The first part of the talk (work with\nMisha Ahrens) involves judgements 
 about short intervals of time.  I\nwill show that an inferential view can 
 help to make sense of\nexperiments where the nature of a stimulus being ti
 med has a\npronounced effect on observers' estimates of its duration\, whi
 le\nnaturally respecting the scalar properties of interval estimates.\nThr
 ee new behavioural experiments support the proposed inferential\nmodel.  I
 n the second part (work with Phillipp Hehrmann) I will\ndiscuss the percep
 t of acoustic pitch.  Again\, an inferential view\nhelps to reconcile a va
 riety of reported behavioural phenomena\, and\nmakes predictions regarding
  the distribution of octave-step errors\nthat are borne out by new experim
 ents.\n\nThus\, even apparently 'simple' percepts may rely on sophisticate
 d\nprocesses of inferential reasoning\, and so may depend on general\nneur
 al mechanisms for pattern recognition and probalistic processing\,\nrather
  than solely on simple dedicated circuits.
LOCATION:Cambridge University Engineering Department\, CBL Rm #438 (http:/
 /learning.eng.cam.ac.uk/Public/Directions)
END:VEVENT
END:VCALENDAR