We will do all we can to make sure that the relationship that you have with The Open University is a positive one.
We will do all we can to treat you as an individual, whether you are contacting to give us some good news or to raise a problem. Please also be assured that we will not share, sell or give your details to anyone.
My research focuses on Natural Language Processing (NLP), and in particular how people interpret ambiguous text. At the Open University, I have investigated how to build computational models that predict when different people will disagree on the interpretation of ambiguous text. My work has looked at both the theoretical foundations of the phenomenon, and its potential impact in the area of requirements engineering.
I am also interested in the problem of how to recognise semantic similarity between texts. My research in this area applies inductive logic programming to learn systems for automatically grading students' written work. This work is currently being extended to use deep learning models for grading, as well as traditional symbolic methods.
The methods underlying these tasks are generalisable beyond NLP. I have collaborated with computational musicologists on machine learning techniques to support automatic music composition for games, and with social scientists on using social media to understand the audience for the Russian television network RT.
I now lead the Artificial Intelligence and Natural Language Processing research group in the School of Computing and Communications, and leads the Data Science theme of the Institute of Coding.
Crilley, Rhys, Gillespie, Marie , Kazakov, Vitaly and Willis, Alistair(2021). ‘Russia isn’t a country of Putins!’: How RT bridged the credibility gap in Russian public diplomacy during the 2018 FIFA World Cup. The British Journal of Politics and International Relations (Early Access).
Crilley, Rhys , Gillespie, Marie , Vidgen, Bertie and Willis, Alistair(2020). Understanding RT’s Audiences: Exposure Not Endorsement for Twitter Followers of Russian State-Sponsored Media. The International Journal of Press/Politics (Early Access).
Amidei, Jacopo , Piwek, Paul and Willis, Alistair(2020). Identifying Annotator Bias: A new IRT-based method for bias identification. In: Proceedings of The 28th International Conference on Computational Linguistics (COLING), 8-13 Dec 2020, Barcelona, Spain, pp. 4787–4797.
Ruiz-Marcos, Germán , Willis, Alistair and Laney, Robin(2020). Automatically calculating tonal tension. In: The 2020 Joint Conference on AI Music Creativity (Sturm, Bob and Elmsley, Andy eds.), 19-23 Oct 2020, Royal Institute of Technology (KTH), Stockholm, Sweden [On-line].
Mensio, Martino,Alani, Harith and Willis, Alistair(2020). Towards a Cross-article Narrative Comparison of News. In: Proceedings of the Text2Story’20 Workshop (Campos, Ricardo; Jorge, Alípio; Jatowt, Adam and Bhatia, Sumit eds.), CEUR WS.
Amidei, Jacopo , Piwek, Paul and Willis, Alistair(2019). Agreement is overrated: A plea for correlation to assess human evaluation reliability. In: Proceedings of the 12th International Conference on Natural Language Generation, 29 Oct - 1 Nov 2019, Tokyo, Japan.
Amidei, Jacopo , Piwek, Paul and Willis, Alistair(2019). The use of rating and Likert scales in Natural Language Generation human evaluation tasks: A review and some recommendations. In: Proceedings of the 12th International Conference on Natural Language Generation, 29 Oct - 1 Nov 2019, Tokyo, Japan.
Crilley, Rhys , Gillespie, Marie and Willis, Alistair(2020). Tweeting the Russian revolution: RT’s #1917LIVE and social media re-enactments as public diplomacy. European Journal of Cultural Studies, 23(3) pp. 354–373.
Amidei, Jacopo , Piwek, Paul and Willis, Alistair(2018). Evaluation methodologies in Automatic Question Generation 2013-2018. In: Proceedings of The 11th International Natural Language Generation Conference, 5-8 Nov 2018, Tilburg, The Netherlands, pp. 307–317.