BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Date iCal//NONSGML kigkonsult.se iCalcreator 2.18//
METHOD:PUBLISH
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
BEGIN:STANDARD
DTSTART:20191103T020000
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20200308T020000
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:calendar.181.field_slac_event_date.0@ml.slac.stanford.edu
DTSTAMP:20200715T083154Z
CREATED:20190923T224906Z
DESCRIPTION:Probabilistic generative models are robust to noise\, uncover u
nseen patterns\, and make predictions about the future. Probabilistic gene
rative models posit hidden structure to describe data. They have addressed
problems in neuroscience\, astrophysics\, genetics\, and medicine. The ma
in computational challenge is computing the hidden structure given the dat
a --- posterior inference. For most models of interest\, computing the pos
terior distribution requires approximations like variational inference. Cl
assically\, variational inference was feasible to deploy in only a small f
raction of models. We develop black box variational inference. Black box v
ariational inference is a variational inference algorithm that is easy to
deploy on a broad class of models and has already found use in neuroscienc
e and healthcare. The ideas around black box variational inference also fa
cilitate new kinds of variational methods such as hierarchical variational
models. Hierarchical variational models improve the approximation quality
of variational inference by building higher-fidelity approximations from
coarser ones. Black box variational inference opens the doors to new model
s and better posterior approximations. Lastly\, I will discuss some recent
generic techniques in finding important features in predictive models lik
e neural networks or random forests.
DTSTART;TZID=America/Los_Angeles:20190926T150000
DTEND;TZID=America/Los_Angeles:20190926T163000
LAST-MODIFIED:20190923T224906Z
SUMMARY:Black Box Variational Inference: Scalable\, Generic Bayesian Comput
ation and its Applications
URL;TYPE=URI:https://ml.slac.stanford.edu/events/black-box-variational-infe
rence-scalable-generic-bayesian-computation-and-its-applications
END:VEVENT
END:VCALENDAR