[Forum SIS] seminar @Chalmers/GU: "Introducing Zig-Zag Sampling and making it applicable", Alice Corbella, 19 October
Umberto Picchini
umberto.picchini a gmail.com
Sab 16 Ott 2021 09:01:03 CEST
You are welcome to the next Statistics seminar at Dept of Mathematical
Sciences at Chalmers and Göteborg University.
We are very glad to have, on Tuesday 19 October,Alice Corbella
(Warwick Uni.) who will talk of:
/"//Introducing Zig-Zag Sampling and making it applicable"/
*//*
*Abstract *
Recent research showed that Piecewise Deterministic Markov Processes
(PDMP) may be exploited to design efficient MCMC algorithms [1]. The
Zig-Zag sampler is an example of this: it is based on the simulation of
a PDMP whose switching rate /λ/(t) is governed by the derivative of a
(minus log) target density.
While many theoretical properties of this sampler have been derived,
less has been done to explore the applicability of the Zig-Zag sampler
to solve Bayesian inference problems. In particular, the computation of
the derivative of the log-density in the rate /λ/(t) might be
challenging. To expand the applicability of the Zig-Zag sampler, we
incorporate Automatic Differentiation tools in the Zig-Zag algorithm, to
evaluate /λ/(t) from the functional form of the log-target density.
Moreover, to allow the simulation of a PDMP via Poisson thinning, we use
univariate optimization routines to find local upper bounds.
In this talk we introduce PDMPs and the Zig-Zag sampler; we expose our
Automatic Zig-Zag sampler; we discuss the challenges that arise with the
simulation via thinning and the need of a new tuning parameter; and we
comment on efficiencies and bottlenecks of AD for Zig-Zag. We present
many examples to compare our method to HMC, another widely used
gradient-based method.
This is joint work with Simon Spencer and Gareth Roberts.
[1] Fearnhead, P., Bierkens, J., Pollock, M., and Roberts, G.O., 2018.
Piecewise deterministic Markov processes for continuous-time Monte
Carlo. Statistical Science, 33(3), pp.386-412. *
*
*
*Feel free to spread this announcement in your network.*
*
*
*Where*: *room PASCAL and zoom* (physical attendance in Pascal is very
welcome)
https://chalmers.zoom.us/j/63909697211
Password: 219826
**
*When*: Tuesday 19 October at 14.15-15.15 CEST (= Swedish time).
About the speaker
Corbella
Alice Corbella is a Post-Doctoral Research Associate of the Bayes4Health
<https://www.lancaster.ac.uk/newbayes/> collaboration, a project that
aims at developing new advanced methods to tackle health problems.
Her current interests are MCMC methods, particularly focussing on the
Zig-Zag sampling and its application to Bayesian inference of infectious
disease dynamics from epidemic data.
Her webpage
<https://warwick.ac.uk/fac/sci/statistics/staff/academic-research/corbella/>.
**
*
--
_________________________________________________________
Umberto Picchini, Associate Professor, PhD, Docent
https://umbertopicchini.github.io/ , twitter: @uPicchini
-------------- parte successiva --------------
Un allegato HTML è stato rimosso...
URL: <http://www.stat.unipg.it/pipermail/sis/attachments/20211016/1d91f0d3/attachment.html>
-------------- parte successiva --------------
Un allegato non testuale è stato rimosso....
Nome: corbella.jpg
Tipo: image/jpeg
Dimensione: 121801 bytes
Descrizione: non disponibile
URL: <http://www.stat.unipg.it/pipermail/sis/attachments/20211016/1d91f0d3/attachment.jpg>
Maggiori informazioni sulla lista
Sis