BEGIN:VCALENDAR
VERSION:2.0
PRODID:talks.ox.ac.uk
BEGIN:VEVENT
SUMMARY:Likelihood-free inference for time-series simulation models - Joel
Dyer (Complexity Economics\, INET Oxford)
DTSTART;VALUE=DATE-TIME:20211104T150000Z
DTEND;VALUE=DATE-TIME:20211104T160000Z
UID:https://talks.ox.ac.uk/talks/id/a1fb946d-7c22-4b8f-8fae-9988d011c356/
DESCRIPTION:Agent-based models often lack a tractable likelihood function\
, precluding classical likelihood-based statistical inference and paramete
r estimation. As a result\, likelihood-free approaches have emerged in rec
ent decades as a means to performing statistical inference for such models
. An example is approximate Bayesian computation\, in which the pertinence
of parameter settings is assessed by some meaningful notion of distance b
etween the simulated and observed data. Agent-based models present a parti
cular challenge in this respect: they often generate time-series data\, wh
ich can be high-dimensional and complex in structure. In this talk\, we wi
ll discuss recently developed approaches to performing Bayesian inference
for intractable time-series simulation models. We will first discuss the p
roblem of likelihood-free inference for time-series simulation models\, be
fore discussing some existing approaches that are popular in the literatur
e. We will then discuss recent work on the use of path signatures as a mea
ns to deriving approximate Bayesian posteriors for time-series simulators
(preprint available at https://arxiv.org/abs/2106.12555)\, and discuss the
properties of path signatures that make them appealing tools in this cont
ext. We will then present experimental results on their use in approximate
Bayesian computation and\, if time allows\, we will discuss ongoing work
on parameter inference for economic agent-based models.\nSpeakers:\nJoel D
yer (Complexity Economics\, INET Oxford)
LOCATION:Virtual Seminar via Zoom (please register to attend)
TZID:Europe/London
URL:https://talks.ox.ac.uk/talks/id/a1fb946d-7c22-4b8f-8fae-9988d011c356/
BEGIN:VALARM
ACTION:display
DESCRIPTION:Talk:Likelihood-free inference for time-series simulation mode
ls - Joel Dyer (Complexity Economics\, INET Oxford)
TRIGGER:-PT1H
END:VALARM
END:VEVENT
END:VCALENDAR