BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:CCN Seminar with Magnus Tournoy (University of Chicago)
DTSTART;VALUE=DATE-TIME:20221121T153000Z
DTEND;VALUE=DATE-TIME:20221121T173000Z
DTSTAMP;VALUE=DATE-TIME:20230206T165600Z
UID:indico-event-3451@indico.flatironinstitute.org
DESCRIPTION:Please join us for a CCN Seminar with Magnus Tournoy\, Postdoc
toral Fellow at the University of Chicago and a candidate for an FRF posit
ion. To schedule a meeting with Magnus during his visit\, please be in t
ouch with Jessica Hauser (jhauser@flatironinstitute.org). Title: A Toy
Model for Uncovering The Structure of Nonlinear Neural NetworksAbstract:Wi
th the experimental advances in the recording of large populations of neur
ons\, theorists are in the humbling position of making sense of a staggeri
ng amount of data. One question that will become more into reach is how ne
twork structure relates to function. But going beyond explanatory models a
nd becoming more predictive will require a fundamental approach. In this
talk we'll take the view of a physicist and formulate exact results within
a simple\, yet general\, toy model called Glass networks. Named after its
originator Leon Glass\, they are the infinite gain limit of well-known ci
rcuit models like continuous-time Hopfield networks. We'll show that\, wit
hin this limit\, stability conditions reduce to semipositivity constraints
on the synaptic weight matrix. Having a clear link between structure and
function in possession\, the consequences of multistability on the networ
k architecture can be explored. One finding is the factorization of the we
ight matrix in terms of nonnegative matrices. Interestingly this factoriza
tion completely identifies the existence of stable states. Another result
is the reduction of allowed sign patterns for the connections. A consequen
ce hereof are lower bounds on the number of excitatory and inhibitory conn
ections. At last we will discuss the special case of "sign stability"\, wh
ere stability is guaranteed by the topology of the network. Derivations of
these results will be supplemented by a number of examples. \n\nhttps:
//indico.flatironinstitute.org/event/3451/
LOCATION:4th Floor Classroom/4-Simons Foundation (160 5th Avenue)
URL:https://indico.flatironinstitute.org/event/3451/
END:VEVENT
END:VCALENDAR