Mathematics for Intelligent Systems

Machine learning (ML) has consistently been the great triumph of pattern recognition ever since the deep learning era begun in 2012. Indeed, the scale of investment in ML has been phenomenal during the past few years. As well as the increased data availability and hardware power, the startling achievements by ML approaches also predominantly rely on intelligent mathematical models.

The Mathematics for Intelligent Systems research group focuses on the development of data-driven mathematical models for the intelligent processing of large amounts of data towards:

1. automating intellectual tasks normally performed by humans,
2. learning efficient (dense) data representations,
3. revealing hidden patterns in the data,
4. optimising decision-making. Problems in both science and engineering are addressed.

Academic collaborators:

• Columbia University
• Lawrence Berkeley National Laboratory
• University of Aarhus
• University of Leicester
• University of Nottingham
• University of Manchester
• University of St. Gallen

Industrial collaborators:

• Keen AI Ltd
• Segmentum Analysis Ltd

Department staff

• Archontis Giannakidis

Fixed-term contract staff

• David Jenkins - Academic Associate
• Tuan Bohoran - MSCA Research Fellow

Visiting Researchers

• Michael Lystbaek - University of Aarhus

• Bohoran TA, Kampaktsis PN, McLaughlin L, Leb J, Moustakidis S, McCann GP, Giannakidis A. (2023). Right Ventricular Volume Prediction by Feature Tokenizer Transformer-based Regression of 2D Echocardiography Small-Scale Tabular Data. In the Proceedings of Functional Imaging and Modeling of the Heart (FIMH 2023). Lecture Notes in Computer Science series, Springer. Vol. 13958.
• Siouras A., Moustakidis S., Giannakidis A., Chalatsis G., Malizos K. N., Hantes M., Tasoulis S., Tsaopoulos D. (2022). Automated Recognition of Healthy Anterior Cruciate Ligament in Sagittal MR Images using Lightweight Deep Learning. In Proceedings of the 13th IEEE International Conference on Information, Intelligence, Systems and Applications (IEEE IISA 2022), July 18-20, Corfu, Greece.
• Siouras A., Moustakidis S., Giannakidis A., Chalatsis G., Liampas I., Vlychou M., Hantes M., Tasoulis S., Tsaopoulos D. (2022). Knee injury detection using deep learning on MRI studies: A systematic review. Diagnostics 12(2):537.
• Giannakidis A., Melkus G., Yang G., Gullberg G. T. (2016). On the averaging of cardiac diffusion tensor MRI data: The effect of distance function selection. Physics in Medicine and Biology 61(21):7765--7786.
• Giannakidis A., Kamnitsas K., Spadotto V., Keegan J., Smith G., Glocker B., Rueckert D., Ernst S., Gatzoulis M. A., Pennell D. J., Babu-Narayan S., Firmin D. N. (2016). Fast Fully Automatic Segmentation of the Severely Abnormal Human Right Ventricle from Cardiovascular Magnetic Resonance Images using a Multi-scale 3D Convolutional Neural Network. In Proceedings of the 12th IEEE International Conference on Signal Image Technology & Internet Systems (IEEE SITIS 2016), November 28 - December 1, Naples, Italy, pp. 42--46.
• Giannakidis A., Nyktari E., Keegan J., Pierce I., Suman-Horduna I., Haldar S., Pennell D. J., Mohiaddin R., Wong T., Firmin D. N. (2015). Rapid automatic segmentation of abnormal tissue in late gadolinium enhancement cardiovascular magnetic resonance images for improved management of long-standing persistent atrial fibrillation. BioMedical Engineering OnLine 14:88.
• Giannakidis A., Gullberg G. T. (2011). Tomographic Reconstruction of 3D Cardiac Diffusion Tensor Fields by Utilizing Reduced Number of Projection Measurements. In Proceedings of the 11th International Meeting on Fully Three-Dimensional Image Reconstruction in Radiology and Nuclear Medicine (Fully 3D 2011), July 11-15, Potsdam, Germany, pp. 455--458.
• Giannakidis A., Petrou M. (2010). Sampling bounds for 2-D vector field tomography. Journal of Mathematical Imaging and Vision 37(2), 151--165.

• Fully automated quantification of myocardial infarct size using artificial intelligence methods
• Revolutionising landslide-tsunami prediction with machine learning
• Developing an artificial intelligence application to support autonomous building design
• Designing weight and cost optimised automotive joints for novel lightweight materials using computer-aided engineering
• Towards a physics-informed deep reinforcement learning framework for more accurate large-eddy simulation of turbulent flows
• Reform of the Gender Recognition Act 2004