BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Making connections- brains and other complex syste
ms
SUMMARY:Statistical analysis and optimality of biological
systems - Prof Gasper Tkacik
DTSTART;TZID=Europe/London:20210422T150000
DTEND;TZID=Europe/London:20210422T160000
UID:TALK158668AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/158668
DESCRIPTION:Normative theories and statistical inference provi
de complementary approaches for the study of biolo
gical systems. A normative theory postulates that
organisms have adapted to efficiently solve essent
ial tasks and proceeds to mathematically work out
testable consequences of such optimality\; paramet
ers that maximize the hypothesized organismal func
tion can be derived ab initio\, without reference
to experimental data. In contrast\, statistical in
ference focuses on the efficient utilization of da
ta to learn model parameters\, without reference t
o any a priori notion of biological function. Trad
itionally\, these two approaches were developed in
dependently and applied separately. Here\, we unif
y them in a coherent Bayesian framework that embed
s a normative theory into a family of maximum-entr
opy “optimization priors.” This family defines a s
mooth interpolation between a data-rich inference
regime and a data-limited prediction regime. I wil
l illustrate how this framework can productively g
uide our thinking on several neuroscience and non-
neuroscience examples.
LOCATION:Online
CONTACT:Sarah Morgan
END:VEVENT
END:VCALENDAR