Topic Models within Network Analysis
Dr Iulia Cioroianu
Dr Tom Fincham Haines
No more applications being accepted
Competition Funded PhD Project (European/UK Students Only)
Online communication takes place in increasingly complex environments in which individuals exchange ideas and interact with others in dynamic social networks. The topics of these conversations, as well as their network evolution can reveal important features of the political and social world and have so far been studied mainly through two methods: topic modelling and social network analysis. Topic models  are an unsupervised machine learning technique for summarising the contents of documents. For instance, if run on news articles they can extract topics such as ‘sport’ and ‘politics’. In the context of social networks they may be run on social media status updates to categorise topics of interest to a user or infer user political preferences. Independently, social networks may have their network structure analysed  to determine clusters of users that share similar beliefs, identify influential individuals or study the ways in which covariates change the structure of the network.
However, very limited research has been conducted on integrating these two approaches. Existing work  is limited in its scope and fails to address the challenges posed by the complexity of social media data. The goals of this project are: 1. to explore and innovate around approaches that combine topic models and social network analysis in order to add depth to the analysis; and 2. to develop software packages that allow computational scientists to easily use the newly developed algorithms. Possible social science applications include exploring the role of online information exposure in political polarization or studying the evolution of social media conversations and the ways in which arguments and their persuasive power are shaped by the structure of the social network.
To arrive at supportable conclusions such studies require transparent algorithms, such as graphical models. In particular, Bayesian parametric approaches such as Dirichlet  and Gaussian  processes can support transparency, report reliability and adapt model complexity to match the level justified by the quantity of available data.
Informal enquiries are welcome and should be directed to Dr. Tom SF Haines ([Email Address Removed]) or Dr. Iulia Cioroianu ([Email Address Removed]).
The project is associated with the UKRI CDT in Accountable, Responsible and Transparent AI (ART-AI). Further details of the programme can be found at: http://www.bath.ac.uk/centres-for-doctoral-training/ukri-centre-for-doctoral-training-in-accountable-responsible-and-transparent-ai/.
Applicants should normally have a good first degree or a Master’s degree in computer science, maths, or a related discipline. A strong mathematical background is essential; good programming skill and previous machine learning experience highly desirable.
Formal applications should be made via the University of Bath’s online application form: https://samis.bath.ac.uk/urd/sits.urd/run/siw_ipp_lgn.login?process=siw_ipp_app&code1=RDUCM-FP02&code2=0002. Enquiries about the application process should be sent to [Email Address Removed].
Start date: 28 September 2020.
ART-AI CDT studentships are available on a competition basis for UK and EU students for up to 4 years. Funding will cover UK/EU tuition fees as well as providing maintenance at the UKRI doctoral stipend rate (£15,009 per annum in 2019/20, increased annually in line with the GDP deflator) and a training support fee of £1,000 per annum.
We also welcome all-year-round applications from self-funded candidates and candidates who can source their own funding.
 “Finding scientific topics” by Griffiths & Steyvers, 2004
 “Network Science” by Barabasi, 2016 (book)
 “Social-Network Analysis Using Topic Models” by Kuang, Chae, Hughes & Natriello, 2017. See also http://www.scottbot.net/HIAL/[email protected]=221.html
 “Hierarchical Dirichlet Processes”, by Teh, Jordan, Beal and Blei, 2005
 “Gaussian Processes for Machine Learning”, by Rasmussen & Williams, 2006 (book)
How good is research at University of Bath in Computer Science and Informatics?
FTE Category A staff submitted: 24.00
Research output data provided by the Research Excellence Framework (REF)
Click here to see the results for all UK universities