Dr Adnane Ez-zizi is a Lecturer in Artificial Intelligence and Data Science at the School of Engineering, Arts, Science and Technology. He holds a PhD in Experimental Psychology from the University of Bristol, where he was part of a large interdisciplinary research group that brought together psychologists, mathematicians, computer scientists and biologists to study human and animal decision making. He also obtained an MRes in Statistics from the University of Pierre et Marie Curie (now Sorbonne University), an MSc in Random modelling from the University of Paris Diderot and a BSc in Statistical techniques from the University of Paris Descartes.
Before joining the University of Suffolk, Adnane worked at the University of Birmingham as a Research Associate in Machine learning within another large interdisciplinary project that combines linguistics, psychology, computer science and statistics to better understand language learning and to improve language teaching. He also took a Senior Teaching Associate role at the University of Bristol for one year, where he taught statistics to undergraduate students.
His work is interdisciplinary in nature, where he has resorted to experimental, statistical, machine learning and natural language processing methods to analyse various types of data. A central theme of his research revolves around the development and testing of computational models that describe human learning behaviour. He has been particularly interested in two forms of learning: (1) reward-based learning (aka. reinforcement learning in the machine learning world) and (2) language learning (with heavy use of NLP techniques). What excites him about this line of work is that not only it can explain some of the hidden facets of human cognition and behaviour but also has the potential to inform the development of faster and more robust machine learning algorithms. Adnane is also very keen on developing tools and software that can help researchers and practitioners in humanities and social sciences analyse their data easier and faster, and has, for example, developed two Python packages for language researchers: one to simplify the tasks of processing and modelling language data (Deep Text Modelling), and one to help them syllabify large text corpora based on IPA (ipa_syllabifiers).
Adnane is currently teaching the following modules:
- Introduction to Artificial Intelligence
- Introduction to Artificial Intelligence and Machine learning
- SQL and NoSQL Databases
List of publications:
Ez-zizi, A., Divjak, D. and Milin, P. (Under revision). Error-correction mechanisms in language learning: modeling individuals.
Romain, L., Ez-zizi, A., Milin, P., and Divjak, D. (Under review). What makes the past perfect and the future progressive? Experiential coordinates for a learnable, context-based model of tense and aspect.
Divjak, D., Milin, P., Ez-zizi, A., Józefowski, J. and Adam, C. (2020). What is learned from exposure: an error-driven approach to productivity in language. Language, Cognition and Neuroscience, 36:1, 60-83.
Ez-zizi, A., McNamara, J.M., Malhotra, G. and Houston, A.I. (2018). Optimal gut size of small birds and its dependence on environmental and physiological parameters. Journal of Theoretical Biology, 454, 357-366.
Ez-zizi, A., Farrell, S. and Leslie, D. (2015). Bayesian reinforcement learning in Markovian and non-Markovian tasks. Proceedings of IEEE Symposium Series on Computational Intelligence, pages 579-586, Cape town.