BEGIN:VCALENDAR
PRODID:-//Columba Systems Ltd//NONSGML CPNG/SpringViewer/ICal Output/3.3-
M3//EN
VERSION:2.0
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
DTSTAMP:20230920T152809Z
DTSTART:20231004T133000Z
SUMMARY:Part-A: A brief and selective review of applications of RMT\; Par
t B: Estimation of rectangular random matrices: Applications to high-dim
ensional biological data (- in person)
UID:{http://www.columbasystems.com/customers/uom/gpp/eventid/}z1x9-lmrup7
9q-7txi9r
DESCRIPTION:Madhuchhanda Bhattacharjee\, Reader in Statistics in the Depa
rtment of Mathematics\, University of Manchester.\n\nTitle: Part-A: A b
rief and selective review of applications of RMT\; Part B: Estimation of
rectangular random matrices: Applications to high-dimensional biologica
l data\n\nAbstract: Part-A: A brief and selective review of application
s of RMT\n\nRandom Matrices have been encountered in Statistics\, Mathem
atics\, Physics\, and \nother disciplines for nearly a century now. Lite
rature on the theoretic developments \nin this field has grown substanti
ally since then. With that there have been wide range \nof applications
of such theories. We discuss a few examples involving\, the more \npopul
ar\, square random matrices and a few where the matrix is rectangular.\n
\nPart B: Estimation of rectangular random matrices: Applications to hig
h-dimensional biological data\n\nIt is commonly thought that for a high-
dimensional data one of the key objectives \nwould be to study it at a l
ower dimensional level. Contrary to this\, there are many \nexamples in
real life where it might be necessary to study it at a latent finer (or
further \nhigher dimensional) level. \n\nConsider such an observed (rand
om) vector\, say Y. then the task would be to explain \nY by means of a
collection of random vectors\, say presented in a rectangular random \nm
atrix X\, where X is not observed.\n\nIn the two applications we conside
r here\, we have a partially observed (random) \nbinary matrix\, say B\,
and a completely unobserved (random) “signal” matrix\, say S\, \nsuch t
hat X can be viewed as a (Hadamard) product of B and S.\nOur two example
s are from neural coding and genomics respectively\, where one of \nthe
dimensions of X is in millions or at least hundreds of thousands. The ot
her \ndimension\, in both of our examples\, is unknown. \nIt is evident
that the problem is ill posed and has many feasible solutions. Our first
\nobjective would be to propose a model that captures the desired highe
r dimensional \nunderstanding\, which we do by using domain knowledge as
far as possible. Next is \nto obtain estimate of the random components
and parameters involved. For which \nwe use an Estimation-Maximization (
EM) type algorithm for one and the Bayesian \nhierarchical modelling for
the other example\n\nVenue: G.114\nAlan Turing Building \nManchester\nM
13 9PL
STATUS:TENTATIVE
TRANSP:TRANSPARENT
CLASS:PUBLIC
LOCATION:Frank Adams Seminar Room 2 \, Alan Turing Building \, Upper Broo
k street\, Manchester\, M13 9PL
END:VEVENT
END:VCALENDAR