I am a computer scientist and data scientist. My academic research has focused on using computational methods for analysing the influence of online information and disinformation, using real-world data of online human interactions.
I am currently a research associate of the Unviersity of Oxford, Oxford Internet Institute. My research at Oxford has focused on measuring and predicting the impact and reach of online disinformation using social media data. My PhD research was on causal inference for estimating the social influence of online communications on real-world outcomes, at the individual and at the collective level.
I hold a PhD in Computer Science from the University of Southampton (2018), an MSc in Operational Research from the University of Southampton (2014, Distinction, dissertation prize, full scholarship), and a BA (Hons) in Computer Science from the University of Cambridge (St John's College 2012, 2.1, dissertation high commendation).
My research focus is on developing computational models and methods for understanding and analysing patterns of behaviour in online interactions, informed by the social sciences. My research interests relate to the areas of online social influence, causal inference, social network analysis, computational social science, and data science.
Liotsiou, D., Moreau, L., & Halford, S. (2016) "Social Influence: From Contagion to a Richer Causal Understanding." In International Conference on Social Informatics (pp. 116-132). Springer International Publishing. [paper on SpringerLink, author copy; poster] Best Poster Award for poster accompanying 17-page full-length paper in proceedings, plus short talk.
Howard, P. N., Ganesh, B., Liotsiou, D., Kelly, J., & François, C. (2018) "The IRA, Social Media and Political Polarization in the United States, 2012-2018." Working Paper 2018.2. Oxford, UK: Project on Computational Propaganda. comprop.oii.ox.ac.uk. 46 pp. (Author order: Oxford P.I., then Oxford postdocs alphabetically, then collaborators from Graphika).
Policy impact: This paper was an independent study requested by the US Senate Select Committee on Intelligence (SSCI), and was cited in Volume II of the SSCI's final report on this topic. It was also cited by the UK House of Commons Digital, Culture, Media and Sport (DCMS) Committee in their Final Report on Disinformation and 'fake news' (2019). .
Selected news coverage: MSNBC interview, The Washington Post (cover story above the fold and further articles), The New York Times (cover story above the fold and further articles), PBS News Hour, ABC News, BBC, The Guardian, The Independent, Ars Technica (with interview quotes), Yahoo Finance (with interview quotes) [more details]
Liotsiou, D., Kollanyi, B., and Howard, P. N. (2019) "The Junk News Aggregator: Examining junk news posted on Facebook, starting with the 2018 US Midterm Elections." arXiv preprint arXiv:1901.07920 Press coverage: TechCrunch, Newsweek, BuzzFeed News, the Bulletin of the Atomic Scientists (video interview).
Peer-reviewed conference presentations
- Liotsiou, D. and Howard, P. N. (2019) "Measuring the Influence of Online Misinformation: A Hierarchy of Social Media Data." The 5th Annual International Conference on Computational Social Science (IC2S2). [paper, poster]
- Liotsiou, D., Moreau, L., and Halford, S. (2017) "Social Influence: from Contagion to a Richer Causal Understanding." The 5th annual UK Causal Inference Meeting, University of Essex, United Kingdom.
- Liotsiou, D., Moreau, L., & Halford, S. A Causal Methodological Framework for Conceptualising and Measuring Social Influence in Online Communications Using Observational Data.
- Liotsiou, D., Moreau, L., & Halford, S. Key Limitations of the Contagion Paradigm for Online Social Influence, and how to Address them.
- Liotsiou, D., Kollanyi, B. & Howard, P. Comparing social media Engagement across Traditional News, Online News, and Junk News, in the Context of the 2018 US Midterm Elections.
- Liotsiou D. (2018, October). Measuring the Social Influence of Online Communications at the Individual and Collective Level: A Causal Framework. PhD Thesis. [abstract]
- Liotsiou D. (2014, September). Projecting Dental Care Need in England over the Next 20-30 Years. Masters Thesis. Sponsor Award.
- Liotsiou D. (2012, June). Parallelising Ant Colony Optimisation-based Solutions to the Vehicle Routing Problem in Scala. Undergraduate Thesis. High Commendation.
PC Member for ACM Conference on Hypertext and Social Media (2019), reviwer for EPJ Data Science (2018)
Other Outreach Activities and Links
- Research talk at the annual Oxbrdige Women in Computer Science conference (2019).
- Oxford University podcast on junk news and AI (2018).
- Profile on the Oxford University AI Research directory, including a discussion of the challenge of causal reasoning in AI (2018).
- Testimonial with advice for prospective Computer Science undergraduate students of St John's College at the University of Cambridge (2013).
During my PhD, I demonstrated and/or marked for the following courses and short seminars:
- Java programming labs, undergraduate-level (demonstator)
- Funtional programming in Scheme, undergraduate-level (marker)
- Social network analysis, postgraduate-level (group project mentor)
- Software engineering, undergraduate-level (group project mentor)
- Introduction to Data Science in Python, for secondary-education teachers
- Introduction to Machine Learning in Python, for PhD students
Some good starting points for causal inference can be found below. This is by no means intended to be an exhaustive list, and the resources are in no particular order. I imagine I will be updating it from time to time. Some good resources for getting started on causality, which use Judea Pearl's do-calculus approach to causality:
- Judea Pearl's 1999 IJCAI Award lecture on causality. On his website he has some more recent slides and tutorials too, but this is one of the two resources he recommends starting with.
- Judea Pearl's overview paper on causality. It is very well written and fun to read. This paper is like a short version of his "Causality" book (see below).
- Judea Pearl's Primer book on Causal Inference in Statistics (co-written with Madelyn Glymour and Nicholas P. Jewell). This is a very accessible introduction to causality and how to practically apply causal methods to data.
- Judea Pearl's book on Causality, which is much longer and more technical than the Primer, and very well written.
- Cosma Shalizi at Carnegie Mellon covers causality really well in his notes/book-in-progress "Advanced Data Analysis from an Elementary Point of View", chapters 20, and 24-28.
- The Morgan-Winship book "Counterfactuals and Causal Inference", the 2nd edition. Particularly: chapter 1.1, parts II (especially 2.1-2.5), III (especially 3.1)
- A brief overview of causal inference, focused on health-related applications, from the Causal Inference group at the Centre for Statistical Methodology of the London School of Hygiene and Tropical Medicine.
- Blog posts:
- Materials from University courses:
Some conferences, workshops, meetings and other academic venues on causality are listed below, from 2017-2018, again in no particular order: