[Turing-Southampton] Announcing Turing at Southampton Seminar Series

[Susan Davies]

***apologies if you receive this more than once***

Turing at Southampton is pleased to announce a new Seminar Series – details of the first two seminars taking place in January are below.  All are welcome to attend!

Speaker:              Shotaro Shiba-Funai, Okinawa Institute of Science and Technology

Date:                    Friday 17 January 2020, 12.00 noon

Location:             Building 46/5081

Title:                    Feature extraction by machine learning and its related physics

Abstract:

Unsupervised machine learning is generally useful to extract features of input data. Since it can be regarded as a kind of information compression, some researchers suggest its similarity to coarse-graining and renormalization.

In this talk, we use spin configurations of Ising model as the input data and restricted Boltzmann machine (RBM) as the method of unsupervised learning. Then we look at what kind of features the machine extracts, using our method of “RBM flow”.  ​As a result, we can find an interesting similarity to renormalization and some coincidence with thermodynamics. However, we also discover apparent differences from renormalization and then argue why such phenomena occur by studying the parameter dependence in machine learning.​

***********************************************

Speaker:             James Burridge, University of Portsmouth

Date:                    Wednesday 29 January 2020, 15.00

Location:             Building 54/5025

Title:                    Language Mechanics

Abstract:

Language is one of the primary “signals” emitted by society - a good place to start if we want a mathematical understanding human collective behavior. In this talk I will explore how the tools of statistical physics can be used to model language evolution. First, I will describe a connection between phase ordering in magnetic systems, and the geographical evolution of language. I will show how different universality classes in physical models correspond to alternative hypotheses for the social factors which drive language change, with some models able to predict observed patterns. I will then explore the sound inventories from which languages are built, specifically vowels. Human vowel systems may be shown to spontaneously emerge from a “soup” of diffusing words, modelled as Brownian particles. If we want such diffusive dynamics to accurately fit real linguistic data, then we can again make use of phase ordering models, treating the evolution of sound inventories as an energy minimization process. Finally, I will discuss how a combination of statistical physics and statistical learning might in future allow us to build models with realistic dynamics which fit complex linguistic datasets.

_____________________________________________
Susan Davies
Coordination Manager, Web Science Institute<https://www.southampton.ac.uk/wsi/index.page?>
University Liaison Manager, The Alan Turing Institute<https://www.southampton.ac.uk/wsi/alan-turing-institute/alan-turing-institute.page>
Room 3041, Building 32
Web Science Institute
University of Southampton
Southampton SO17 1BJ

023 8059 3523 | 07768 266464

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ecs.soton.ac.uk/pipermail/turing-southampton/attachments/20200109/4400c7bf/attachment.html