- Carys Harvey (PhD, co-supervised with Bob Coecke). Carys is working on implementing NLP algorithms on quantum hardware.
- Xiaoyu Tong (PhD, co-supervised with Ekaterina Shutova). Xiaoyu’s research looks at metaphor interpretation via paraphrase and inference.
- Junda Liu (PhD, co-supervised with Nirav Ajmeri). Junda’s research is in ethics in multi-agent systems, specifically driverless vehicles.
- Changshuo Wang (MSc Engineering Maths) Changshuo’s research is in extending neural methods for building density matrices.
- Qiming Chen (MSc Engineering Maths) Qiming looks at modelling generic sentences with density matrices.
- Mike Nelhams (MEng Engineering Maths). Mike is looking at using density matrices on larger-scale entailment datasets.
- Ben Lippold (MEng Engineering Maths). Ben is investigating the interaction of metaphor and natural language inference.
Prospective Students: If you are interested in working with me, please do get in touch. Funding for PhD students is available via the Interactive AI CDT and the Quantum Engineering CDT - funding is mostly for UK students, but a portion is reserved for international students. I am also available to supervise MSc projects and welcome contact from students interested in designing their own project.
- James Owers (MSc Data Science, Bristol, 2021). James investigated the ability of density matrices and state of the art sentence encoders to accurately select the correct paraphrase from a novel metaphor paraphrase dataset. Spoiler alert - none of the models were any good!
- Shuhao Zhang (MSc Data Science, Bristol, 2021). Shuhao looked at novel approaches to building word region embeddings.
- Jiaqi Wu (MSc Data Science, Bristol, 2021). Jiaqi investigated the performance of neural density matrices to encode hyponymy.
- Xiaoyu Tong (MSc Brain and Cognitive Sciences, UvA, 2020, co-supervised with Ekaterina Shutova). Xiaoyu built a large natural language metaphor paraphrase dataset and undertook a systematic review of neural methods for metaphor processing in NLP.
- Francois Meyer (MSc AI, UvA, 2020). Francois built a neural model for generating density matrix word embeddings, showing how these embeddings can represent multiple senses of words which are disambiguated in the process of composition.
- Verna Dankers (MSc AI, UvA, 2019, co-supervised with Ekaterina Shutova). Verna developed a multitask learning architecture to jointly detect metaphor and emotion, as well as a heap of other stuff!
- Yaared al-Mehairi (MFoCS, Oxford, 2016, co-supervised with Bob Coecke). Yaared worked on linking the DisCoCat model with Smolensky’s Integrated Connectionist-Symbolic model.
- Dea Bankova (MFoCS, Oxford, 2015, co-supervised with Bob Coecke). Dea worked on the use of density matrix models for hyponymy and entailment.