Author: cssrajt

Congratulations to Awad who successfully defended his thesis entitled “Combined Supervised and Unsupervised Learning to Identify Subclasses of Disease for Better Prediction” in January. Thanks to examiners Jaakko Hollmen and Stasha Lauria.

Biraja successfully defended his thesis, titled “When the Machine Does Now Know Measuring Uncertainty in Deep Learning Models of Medical Image” on 9th September.

Toyah Overton presented and Juliana Branescu discussed her poster at the first in-person conference for a while in Rennes, France.

Congratulations to those with new publications!:

Toyah Overton, Allan Tucker, Tim James and Dimitar Hristozov. dunXai: DO-U-Net for Explainable (Multi-Label) Image Classification, Intelligent Data Analysis Symposium 2022;

Ghoshal, B., Hikmet, F., Pineau, C., Tucker, A. and Lindskog, C. (2021) ‘DeepHistoClass: A novel strategy for confident classification of immunohistochemistry images using Deep Learning‘. Molecular & Cellular Proteomics, 0 (In Press). pp. 1 – 71. ISSN: 1535-9476

Barnaby E. Walker, Allan Tucker and Nicky Nicolson, Harnessing Large-Scale Herbarium Image Datasets Through Representation Learning, Front. Plant Sci., 13 January 2022 | https://doi.org/10.3389/fpls.2021.806407

The IDA- Group gave a talk at the inaugaural workshop by the HDR UK Synthetic Data Special Interest Group:

Congratulations Yani Xue for successfully defending here thesis titled “Effective and Efficient Evolutionary Many-Objective Optimization”

Congratulations to Lilly Yousefi for passing her PhD viva with minor corrections.

Thesis Title: “Opening Artificial intelligence Black Box Models: Disease Prediction and Patient Personalisation Using Hidden Variable Discovery and Dynamic Bayesian Networks“

Lilly is now working at Brunel Biosciences in collaboration with the Turing Institute and UCL.

A collaboration between the Clinical Practice Research Datalink (CPRD), MHRA Medical Devices Division and researchers at Brunel University has led to the creation of two innovative synthetic datasets which will support the development of cutting-edge medical technologies to fight COVID-19 and cardiovascular disease. More detail here.

Congratulations to Erfan Sajjadi and Ben Evans for their succesful submissions to the SOGOOD workship at ECML this year:

Erfan Sajjadi et al. Building Trajectories over Topology with TDA-PTS: An Application in Modelling Temporal Phenotypes of Disease

Ben Evans et al. Reasoning about Neural Network Activations: An Application in Spatial Animal Behaviour from Camera Trap Classifications

Other new publications from the group include:

Biraja Ghoshal & Allan Tucker, On Calibrated Model Uncertainty in Deep Learning, ECML Workshop on Uncertainty in Machine Learning

Juan de Benediti et al. Practical Lessons from Generating Synthetic Healthcare Data with Bayesian Networks, SOGOOD, ECML 2020

Arianna Dagliati et al. Using Topological Data Analysis and Pseudo Time Series to Infer Temporal Phenotypes from Electronic Health Records, AI in Medicine Journal