Machine Learning incorporating expert knowledge
Kerstin Bunte, University of Groningen
Nowadays, most successful machine learning (ML) techniques for the analysis of complex interdisciplinary data use significant amounts of measurements as input to a statistical system. The domain expert knowledge is often only used in data preprocessing. The subsequently trained technique appears as a “black box”, which is difficult to interpret and rarely allows insight into the underlying natural process. Especially in critical domains such as medicine and engineering, the analysis of dynamic data in the form of sequences and me series is often difficult.
Due to natural or cost limitations and ethical considerations data is often irregularly and sparsely sampled and the underlying dynamic system is complex. Therefore, domain experts currently enter a me-consuming and laborious cycle of mechanistic model construction and simulation, often without direct use of the experimental data or the task at hand.
Recently, hybrid approaches were published combining the predictive power of ML and the explanatory power of pharmacokinetic models for model-based clustering (automatically determining groups of responses to medication in a clinical data set) and classification. In this talk I will give an overview of these new emerging strategies and the concept of “Learning in the model space” (LIMS).
Explainable Models for Personalised Integrated Stroke Ultrasound Video Analysis
Costantinos Pattichis, University of Cyprus
Approximately 15% of all strokes are the result of carotid bifurcation atherosclerosis but at present, there is no established set of features that can identify all the potentially unstable and high-risk plaques. The overall objective of this work is to develop an integrated intelligent system for the stratification of stroke risk patients combining different clinical and imaging risk factors based on the morphology and motion characteristics of the ultrasound video images of carotid bifurcation plaques producing moderate to severe stenosis (more than 70%).
Constantinos S. Pattichis is Professor with the Dep. of Computer Science and Director of the Biomedical Engineering Research Centre at the University of Cyprus and Leader of HealthXR Smart, Ubiquitous, and Participatory Technologies for Healthcare Innovation in the CYENS Centre of Excellence. He has 30 years of experience in eHealth and connected health, medical imaging, biosignal analysis, intelligent systems and explainable AI, and more recently in mHealth interventions based on X Reality applications. He has been involved in numerous projects in these areas funded by EU and other bodies, with a total funding managed close to 14 million Euro. He has published 130 journal publications, 236 conference papers, 30 chapters in books and editor of 3 books, 22 journal special issues and 19 conference proceedings in these areas. He is a Fellow of IEEE, IET, International Academy of Medical and Biomedical Engineering (IAMBE) and European Alliance for Medical & Biological Engineering & Science (EAMBES).
Human Face Recognition: Learning from Biological Deep Networks
Massimo Tistarelli, University of Sassari
Face Recognition has been extensively studied as a mean to facilitate man-machine interaction in a variety of different applications. Due to the imaging variabilities and to the complex nature of the face shape and dynamics, analyzing and recognizing human faces from digital images is still a very complex task.
In the last decade deep learning techniques have strongly influenced many aspects of computational vision. Many difficult vision tasks can now be performed by deploying a properly tailored and trained deep network. Oxford University’s VGG-face is possibly the first deep convolutional network designed to perform face recognition, obtaining unsurpassed performance at the time it was firstly proposed. The enthusiasm for deep learning is unfortunately paired by the present lack of a clear understanding of how they work and why they provide such brilliant performance. The same applies to Face Recognition.
Over the last years, several and more complex deep convolutional networks, trained on very large, mainly private, datasets, have been proposed still elevating the performance bar also on quite challenging public databases, such as the Janus IJB-A and IJB-B. Despite of the progress in the development of such networks, and the advance in the learning algorithms, the insight on these networks is still very limited. For this reason, in this talk we analyse the neural architecture of the early stages of the human visual system to devise a biologically-inspired model for face recognition. The aim is not pushing the recognition performance further, but to better understand the representation space produced from a deep network and how it may help explaining the process undergoing a real biological neural architecture.
In this talk we analyse an hybrid model network trying to better understand the role of the different layers, including the retino-cortical mapping simulated by a log-polar image resampling. The following issues will be addressed:
- What is the representation space within a deep convolutional network and how this reflects the organization of the human visual cortex.
- How the retino-cortical mapping, implemented in the human visual system, may impact the representation space, hence improving the classification performance.
- The relevance of peripheral vs foveal vision for face recognition.
Massimo Tistarelli received the Phd in Computer Science and Robotics in 1991 from the University of Genoa. He is Full Professor in Computer Science (with tenure) and director of the Computer Vision Laboratory at the University of Sassari, Italy. Since 1986 he has been involved as project coordinator and task manager in several projects on computer vision and biometrics funded by the European Community.
Since 1994 he has been the director of the Computer Vision Laboratory at the Department of Communication, Computer and Systems Science of the University of Genoa, and now at the University of Sassari, leading several National and European projects on computer vision applications and image-based biometrics.
Prof. Tistarelli is a founding member of the Biosecure Foundation, which includes all major European research centers working in biometrics. His main research interests cover biological and artificial vision (particularly in the area of recognition, three-dimensional reconstruction and dynamic scene analysis), pattern recognition, biometrics, visual sensors, robotic navigation and visuo-motor coordination. He is one of the world-recognized leading researchers in the area of biometrics, especially in the field of face recognition and multimodal fusion. He is coauthor of more than 150 scientific papers in peer reviewed books, conferences and international journals. He is the principal editor for the Springer books “Handbook of Remote Biometrics” and “Handbook of Biometrics for Forensic Science”.
Prof. Tistarelli organized and chaired several world-recognized several scientific events and conferences in the area of Computer Vision and Biometrics, and he has been associate editor for several scientific journals including IEEE Transactions on PAMI, IET Biometrics, Image and Vision Computing and Pattern Recognition Letters.
Since 2003 he is the founding director for the Int.l Summer School on Biometrics (now at the 17th edition – http://biometrics.uniss.it). He is a Fellow member of the IAPR and Senior member of IEEE, and Vice President of the IEEE Biometrics Council.