Vasileios (aka Bill) Lampos
Associate Professor
AI Centre
Department of Computer Science
University College London
I am interested in machine learning (artificial intelligence) methods for modelling time series (see recent preprint and paper). In addition, developing solutions for natural language processing tasks (e.g. using social media activity to predict voting intention, user impact or socio-economic attributes) and health-related problems (e.g. modelling the prevalence of COVID-19 or influenza-like illness, predicting the impact of a health intervention, and transferring disease models from one country to another) has consistently been the primary focus of my research.
Join us!
Twitter: @lampos
UCL profile
Email: v.lampos (at) ucl.ac.uk
Research news
- *** Join my research group! Opportunities are listed here. ***
- 28/08/2023 — Our paper titled "Neural network models for influenza forecasting with associated uncertainty using Web search activity trends" is now published in PLOS Computational Biology.
- 01/09/2021 — Our paper titled "An artificial intelligence approach for selecting effective teacher communication strategies in autism education" is now published in Nature (npj) Science of Learning.
- 08/02/2021 — Our paper titled "Tracking COVID-19 using online search" is now published in Nature (npj) Digital Medicine. You can find more about our work in this press release.
- 10/09/2020 — A Google.org initiative is supporting our research on COVID-19.
- 18/09/2019 — Our research on transfer learning for disease surveillance models from online search activity was covered by Nature as part of an outlook article about real-time flu tracking.
- 21/01/2019 — 2 papers have been accepted by the Web Conference 2019. The first one proposes a transfer learning method for estimating flu rates using web search activity in locations that do not have an established health surveillance system, and the second proposes a privacy-preserving framework for collecting web search activity data for health-related research.
- 24/05/2018 — The 2017/18 annual flu report by Public Health England incorporates internet-based flu rate estimates, powered by web search activity and our machine learning models.
- 22/12/2017 — Our paper on multi-task learning models for syndromic surveillance from Google search data has been accepted by WWW 2018.
- 01/01/2017 — Our paper proposing a better feature selection method for syndromic surveillance models from web search activity was accepted by WWW 2017.
- 24/10/2016 — Our paper on predicting judicial decisions of the European Court of Human Rights using statistical NLP was published in PeerJ Computer Science; this research has been covered by mainstream media outlets (e.g. VICE, BBC, The Guardian) as well as a UCL press release.