Alina Bialkowski

Dr Alina Bialkowski is a computer vision & machine learning researcher developing interpretable machine learning models to increase the performance and transparency of Artificial Intelligence (AI) decision-making. Her research interests include quantifying and extracting actionable knowledge from data to solve real-world problems and giving human understanding to AI models through feature visualisation and attribution methods. She has applied these techniques to various multi-disciplinary applications such as medical imaging (including imaging strokes in the brain using the new sensing modality of electromagnetic imaging), modelling human attention in driving, intelligent transport systems (ITS), intelligent surveillance, and sports analytics. Dr Bialkowski holds a PhD and BEng (Electrical Engineering) from the Queensland University of Technology, Australia. Her doctoral studies were in characterising group behaviours from visual and spatio-temporal data to enhance statistics and visualisation in sports analytics as well as intelligent surveillance systems. She spent a year at Disney Research Pittsburgh where she developed techniques to automatically analyse team sports, followed by 2.5 years as a postdoctoral researcher at the University College London, developing deep neural networks to better understand human perception and attention in driving, before joining UQ in late 2017. The impact of her research is evidenced by the high number of citations to her work (>1600 citations and an h-index of 20 according to Google Scholar) and awards including a best paper prize in 2017 at WACV (a top computer vision conference). In addition to high impact journals and conferences, her work has resulted in 6 international patents filed with Disney Research, Toyota Motor Europe, University College London, and The University of Queensland. Machine Learning: Developing interpretable or explainable models to increase trust and transparency of computer-based decisions. Computer Vision: Developing models of visual information using image processing and feature representation learning approaches such as deep learning to perform predictions on data. Sensors: Utilising sensors such as cameras and antennas/electromagnetic sensors to enable remote sensing and non-invasive imaging.

[2026] Journal Article: Synthetic expansion of blood dielectric spectra at microwave frequencies using data-driven methods. Alhummada, Iman, Bialkowski, Alina, Guo, Lei, Villena Gonzales, Wilbert, Deriche, Mohamed and Abbosh, Amin (2026). Sensors, 26 (11) 3580, 1-19. doi: 10.3390/s26113580 [2026] Journal Article: Medical microwave imaging using physics-guided deep learning part 2: the inverse solver. Guo, L., Bialkowski, A. and Abbosh, A. (2026). IEEE Transactions on Medical Imaging, 45 (5), 2295-2305. doi: 10.1109/tmi.2026.3653974 [2026] Conference Publication: Deep Learning with Multi-Block Pyramid Sparse Convolution-Attention Network for Enhanced Microwave Imaging. Xue, Fei, Guo, Lei, Bialkowski, Alina and Abbosh, Amin (2026). IEEE. doi: 10.1109/asa69049.2026.11466990 [2026] Journal Art