[Forum SIS] seminar: "Sequential Monte Carlo for Approximate Bayesian Inference", Fredrik Lindsten, 16 March

[Umberto Picchini]

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 16nd March, Fredrik Lindsten 
<https://lindsten.netlify.app/> (Linköping University) who will talk on

/*Sequential Monte Carlo for Approximate Bayesian Inference*/
*//*

Zoom: https://chalmers.zoom.us/j/67555083773
Password: 904889

When: 14.15-15.15 CET, 16 March

Feel free to circulate this invitation in your network.


*​Abstract *
*Sequential Monte Carlo (SMC) is a powerful class of methods for 
approximate Bayesian inference. While originally used mainly for signal 
processing and inference in dynamical systems, these methods are in fact 
much more general and can be used to solve many challenging problems in 
Bayesian statistics and machine learning, even if they lack apparent 
sequential structure. In this talk I will first discuss the foundations 
of SMC from a machine learning perspective. We will see that there are 
two main design choices of SMC: the proposal distribution and the 
so-called intermediate target distributions, where the latter is often 
overlooked in practice. Focusing on graphical model inference, I will 
then show how deterministic approximations, such as variational 
inference and expectation propagation, can be used to approximate the 
optimal intermediate target distributions. The resulting algorithm can 
be viewed as a post-correction of the biases associated with these 
deterministic approximations. Numerical results show improvements over 
the baseline deterministic methods as well as over "plain" SMC.

The first part of the talk is an introduction to SMC inspired by our 
recent Foundations and Trends tutorial
https://www.nowpublishers.com/article/Details/MAL-074 
<https://eur01.safelinks.protection.outlook.com/?url=https://www.nowpublishers.com/article/Details/MAL-074&data=04%7c01%7cfredrik.lindsten%40liu.se%7ce2994ccd25964140a86708d8dee80201%7c913f18ec7f264c5fa816784fe9a58edd%7c0%7c0%7c637504435330479755%7cUnknown%7cTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7c1000&sdata=7DuonzSawLkrDojhD05a5rcHYFSHmFjM7gMcwQ2X3Is%3D&reserved=0>
https://arxiv.org/abs/1903.04797 
<https://eur01.safelinks.protection.outlook.com/?url=https://arxiv.org/abs/1903.04797&data=04%7c01%7cfredrik.lindsten%40liu.se%7ce2994ccd25964140a86708d8dee80201%7c913f18ec7f264c5fa816784fe9a58edd%7c0%7c0%7c637504435330489749%7cUnknown%7cTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7c1000&sdata=X/NYyN/PNAvXAsnUykHZ9v%2Bw4s8mulh3Z43bXxYCehQ%3D&reserved=0>

The second part of the talk, focusing on combining SMC and deterministic 
approximations for graphical model inference, is based on 
https://papers.nips.cc/paper/2018/hash/351869bde8b9d6ad1e3090bd173f600d-Abstract.html 
<https://eur01.safelinks.protection.outlook.com/?url=https://papers.nips.cc/paper/2018/hash/351869bde8b9d6ad1e3090bd173f600d-Abstract.html&data=04%7c01%7cfredrik.lindsten%40liu.se%7ce2994ccd25964140a86708d8dee80201%7c913f18ec7f264c5fa816784fe9a58edd%7c0%7c0%7c637504435330499752%7cUnknown%7cTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7c1000&sdata=0fLx/n07cMLFd8ha8aXVQ7VmsTL90RbvznqBTMV1gFI%3D&reserved=0>


About the speaker
Fredrik Lindsten**

Fredrik Lindsten is associate professor at Linköping University. He 
isinterested in the interplay between statistics and machine learning, 
in particular how statistical methodology can be used to quantify and 
reason about the uncertainties in the predictions and decisions made by 
machine learning systems.

He has received the Ingvar Carlsson Award by the Swedish Foundation for 
Strategic Research, and the Benzelius Award by the Royal Society of 
Sciences in Uppsala.







Welcome!
*

-- 
_________________________________________________________
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/20210312/abf77314/attachment.html>
-------------- parte successiva --------------
Un allegato non testuale è stato rimosso....
Nome:        employee_image_freli29.jpeg
Tipo:        image/jpeg
Dimensione:  10397 bytes
Descrizione: non disponibile
URL:         <http://www.stat.unipg.it/pipermail/sis/attachments/20210312/abf77314/attachment.jpeg>