Work on Mental Health


Over recent years, there has been an increased focus on how advances in NLP can be used to aid mental health research. In our work, we predominantly focus on work in the areas of depression, suicide note and suicide ideation detection, using deep learning and knowledge-based approaches.


Automating Mental Health Interventions for Long COVID funded by the Institute for Data Science & Artificial Intelligence (IDSAI) and the Alan Turing Institute

Building a Mental Health Knowledge Graph funded by the Institute for Data Science & Artificial Intelligence (IDSAI) and the Alan Turing Institute


ICML 2021 Workshop on Computational Approaches to Mental Health - 24th July 2021

NacTeM PhD students Jake Vasilakes and Tianlin Zhang are co-organisers of the ICML 2021 Workshop on Computational Approaches to Mental Health. The workshop aims to bring together clinicians, behavioural scientists and machine learning researchers working in various facets of mental health and care provision, to identify key opportunities and challenges in developing solutions for this domain, and discuss the progress made.

Participation in shared task: e-Risk 2021: Early risk prediction on the Internet

NacTeM members Hassan Alhuzali, Tianlin Zhang and Sophia Ananiadou are participating in CLEF e-risk Task 3, focused on detecting early sign of depression in Reddit data. The growing interest in building effective approaches to this problem has been motivated by the proliferation of social media data, which have made it possible for people to communicate and share opinions on a variety of topics. Our approach is based on pre-trained models plus a standard machine learning algorithm. We specifically use the pre-trained models to extract features for all users' posts and then feed them into a random forest classifier, achieving an average hit rate of 32.86%. With this performance, the system developed at NaCTeM is ranked amonst the 5 top-performing systems that wee submitted to the shared task. The workshop associated with the shared task will take place in Bucharest, from 21-24 September 2021.

Related Publications

Alhuzali, H., Zhang. T. and Ananiadou, S. (In Press). Predicting Sign of Depression via Using Frozen Pre-trained Models and Random Forest Classifier. In CLEF (Working Notes).

Alhuzali, H., and Ananiadou, S. (2021). SpanEmo: Casting Multi-label Emotion Classification as Span-prediction. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume (pp. 1573-1584).

Schoene, A.M., Turner, A., De Mel, G.R. and Dethlefs, N. (2021). Hierarchical Multiscale Recurrent Neural Networks for Detecting Suicide Notes. IEEE Transactions on Affective Computing.

Alhuzali, H., and Ananiadou, S. (2019). Improving classification of adverse drug reactions through using sentiment analysis and transfer learning. In Proceedings of the 18th BioNLP Workshop and Shared Task (pp. 339-347).

Schoene, A.M., Lacey, G., Turner, A.P. and Dethlefs, N. (2019) Dilated LSTM with attention for classification of suicide notes. In Proceedings of the Tenth International Workshop on Health Text Mining and Information Analysis (LOUHI 2019) (pp. 136-145).

Korkontzelos, I., Nikfarjam, A., Shardlow, M., Sarker, A., Ananiadou, S. and Gonzalez, G. (2016). Analysis of the effect of sentiment analysis on extracting adverse drug reactions from tweets and forum posts. Journal of Biomedical Informatics, 62(148-158)

Schoene, A.M. and Dethlefs, N. (2016) Automatic identification of suicide notes from linguistic and sentiment features. In Proceedings of the 10th SIGHUM Workshop on Language Technology for Cultural Heritage, Social Sciences, and Humanities (pp. 128-133).


NacTeM members working on mental health:
Sophia Ananiadou, Annika Schoene, Hassan Alhuzali, Jake Vasilakes, Tianlin Zhang