Maria Liakata

Turing AI Fellow Maria Liakata’s work is creating tools that can help individuals and clinicians monitor mental health conditions

Tell me about your work in a nutshell?

My work focuses on Natural Language Processing (NLP). This is the field of Artificial Intelligence (AI) that is developing computational methods that can process human language to perform different tasks. These tasks could include summarising documents automatically or identifying the accuracy of information on social media.

Specifically, my team at Queen Mary University of London and the Turing are developing NLP methods to identify and understand changes in individuals' behaviour. This behaviour  is demonstrated through their language and other content they produce while interacting with others online or by using different devices.

What do you anticipate will the potential impact of your research?

We hope to create time-sensitive sensors from language and other heterogeneous digital content created by individuals. This will help create tools that can be used by individuals, as well as clinicians, to monitor mental health conditions.

What’s been the most surprising thing to come out of your work?

How hard it is to work with mental health data. The challenges involved range from the size of datasets and the data collection process to how and whom you can share data and models with. As AI researchers we need to learn how to work with sensitive data so we can improve our understanding and monitoring of mental health conditions.

What’s been a career highlight so far?

Being awarded a prestigious EPSRC/UKRI Turing AI Fellowship and becoming a full Professor of Natural Language Processing at Queen Mary University of London.

You’re an organiser of the Turing’s Data Science for Mental Health interest group – can you tell me a bit about this group?

The Data Science for Mental Health Special Interest Group was set up to bring people from different disciplines (such as Maths, Psychiatry, Machine Learning) and backgrounds (academic, industry, general public) together to discuss topics on mental health research. We are motivated by a vision to develop data science methods that can provide a uniform approach to analysing and addressing a range of different mental health conditions.

And finally, when not working what can you be found doing?

Spending time with my four year old. Raising children is a wonderful and formidable project - no matter how many books you read on the subject and how well you know the theory. I try to sneak in some time for exercise (mainly walking and Pilates) to counterbalance all the hours spent sitting in front of my laptop and to help me keep up with my child's energy.