Home      Log In      Contacts      FAQs      INSTICC Portal
 

Keynote Lectures

Visualizing health data – from fundamental research to successful applications
Roy Ruddle, University of Leeds, United Kingdom

Patient Innovation - When Patients Innovate and Improve Their Lives
Helena Canhão, Universidade NOVA de Lisboa, Portugal

Uncertainty Modeling and Deep Learning Applied to Food Image Analysis
Petia Radeva, Universitat de Barcelona, Spain

Towards Robust Machine Learning in the Medical Domain
Andreas Holzinger, Medical University Graz, Austria

 

Visualizing health data – from fundamental research to successful applications

Roy Ruddle
University of Leeds
United Kingdom
 

Brief Bio
Roy Ruddle is a Professor of Computing at the University of Leeds (UK), and an Alexander von Humboldt Foundation and Alan Turing Institute Fellow. He has worked in academia and industry, and researches visualization, visual analytics and human-computer interaction in spaces that range from high-dimensional data to virtual reality (https://raruddle.wordpress.com/).


Abstract
Health data comes in many forms, from relational databases of electronic health records, to genomic sequencing and the tera-pixel collections of images that pathologists use to diagnose diseases such as cancer. This talk will take you on a journey across all of those modalities, connecting fundamental research with successful applications. One of those applications allows users to investigate and explain data quality problems that involve dozens of variables, and others exploit the gigantic display real estate of Powerwalls to speed up research and diagnosis of cancer (Orchestral and the Leeds Virtual Microscope, respectively). Most people think that visualization is only needed to analyse data and present findings. However, in the big data era the most important use of visualization is for designing models and data processing pipelines, so I will conclude with examples involving AI.



 

 

Patient Innovation - When Patients Innovate and Improve Their Lives

Helena Canhão
Universidade NOVA de Lisboa
Portugal
 

Brief Bio

Graduate, PhD and Habillitation in Medicine from Faculty of Medicine, University of Lisbon, Portugal. She also holds a Master's Degree in Clinical Research from Harvard Medical School, Harvard University, USA.

Head of EpiDoC Research Unit, CEDOC, CHRC, NOVA Medical School, Universidade NOVA de Lisboa, Lisbon, Portugal.

She is Full Professor of Medicine at NOVA Medical School, Invited Full Professor at National School of Public Health and Head of Rheumatology Unit, CHLC- Hospital Curry Cabral, Lisbon.

She is board member of the Council of Universidade NOVA de Lisboa, and board member Council of NOVA Medical School, UNL.

Elected President of the Portuguese Society of Rheumatology, President of the EpiSaude Association and Co-leader and Chief Medical Officer for the Patient Innovation Project.

She chairs the Advisory Board of Value4Health CoLab and coordinates NOVA Saude Ageing Group. 

She authored and co-authored more than 200 peer-reviewed publications and edited 5 books and 25 book chapters in the fields of medicine, rheumatology, innovation and clinical research.


Abstract
User-innovation is a new field of research and interest. Users often innovate and develop new solutions in a variety of fields. However, health with all the issues involving ethics, regulations, safety and scientific knowledge would not be an easy field for patients or informal caregivers innovate. But we found that, in fact, patients innovate, some becoming even entrepreneurs, however the diffusion of their solutions is low. Our open, non-profit Patient Innovation platform shares around 1000 solutions developed by patients and caregivers, from simple to high-tech solutions. The solutions are published only after medical screening and each one has a story behind.



 

 

Uncertainty Modeling and Deep Learning Applied to Food Image Analysis

Petia Radeva
Universitat de Barcelona
Spain
 

Brief Bio
Prof. Petia Radeva completed her undergraduate study on Applied Mathematics and Informatics at the University of Sofia, Bulgaria, in 1989. In 1996, she received a Ph.D. degree in Computer Vision at UAB. In 2007, she moved as Tenured Associate professor at the Universitat de Barcelona (UB), Department of Mathematics and Informatics, where from 2009 to 2013 she was Director of Computer Science Undergraduate Studies. In 2018, Petia Radeva became Full professor at the Universitat de Barcelona. Petia Radeva is Head of the Consolidated Group Computer Vision at the University of Barcelona (CVUB) at UB (www.ub.edu/cvub) and Head of the Medical Imaging Laboratory of Computer Vision Center (www.cvc.uab.es). Petia Radeva’s research interests are on Development of learning-based approaches (especially, deep learning) for computer vision, and their application to health. Currently, she is involved on projects that study the application of wearable cameras and life-logging, to extract visual diary of individuals to be used for memory reinforcement of patients with mental diseases (e.g. Mild cognitive impairment). Moreover, she is exploring how to extract semantically meaningful events that characterize lifestyle and healthy habits of people from egocentric data. Petia Radeva is Principal Investigator of the Universitat de Barcelona in two European projects devoted to food intake monitoring for patients with kidney transplants and for older people. There her team applies most recent and advanced methods for food image analysis using Deep learning models. Petia Radeva is REA-FET-OPEN vice-chair from 2015 and has actively participated from 2018 as a mentor in the Wild Cards EIT program. Petia Radeva is associate editor of Pattern Recognition journal and International Journal of Visual Communication and Image Representation. She obtained the ICREA award from the Catalonian Government for her scientific merits in 2014, the international award “Aurora Pons Porrata” from CIARP in 2016 and the Prize “Antonio Caparrós” for the best technology transfer project of 2013.


Abstract
Recently, computer vision approaches specially assisted by deep learning techniques have shown unexpected advancements that practically solve problems that never have been imagined to be automatized like face recognition or automated driving. However, food image recognition has received a little effort in the Computer Vision community. In this project, we review the field of food image analysis and focus on how to combine with two challenging research lines: deep learning and  uncertainty modeling. After discussing our methodology to advance in this direction, we comment potential research, social and economic impact of the research on food image analysis.



 

 

Towards Robust Machine Learning in the Medical Domain

Andreas Holzinger
Medical University Graz
Austria
 

Brief Bio
Andreas Holzinger is lead of the Holzinger Group (Human-Centered AI) at the Medical University Graz and Visiting Professor for explainable AI at the Alberta Machine Intelligence Institute in Edmonton, Canada. Since 2016 he is Visiting Professor for Machine learning in health informatics at Vienna University of Technology. Andreas was Visiting Professor for Machine Learning & Knowledge Extraction in Verona, RWTH Aachen, University College London and Middlesex University London. He serves as consultant for the Canadian, US, UK, Swiss, French, Italian and Dutch governments, for the German Excellence Initiative, and as national expert in the European Commission. Andreas obtained a Ph.D. in Cognitive Science from Graz University in 1998 and his Habilitation in Computer Science from TU Graz in 2003. He founded the Network HCI-KDD to foster a synergistic combination of methodologies of two areas that offer ideal conditions toward unraveling problems in understanding intelligence towards context-adaptive systems: Human-Computer Interaction (HCI) & Knowledge Discovery/Data Mining (KDD), with the goal of supporting human intelligence with artificial intelligence. Andreas is Associate Editor of Springer/Nature Knowledge and Information Systems (KAIS), Section Editor for Machine Learning of Springer/Nature BMC Medical Informatics and Decision Making (MIDM), and Editor-in-Chief of Machine Learning & Knowledge Extraction (MAKE). In his function as Austrian Representative for Artificial Intelligence in IFIP TC 12, he is organizer of the IFIP Cross-Domain Conference “Machine Learning & Knowledge Extraction (CD-MAKE)” and is member of IFIP WG 12.9 Computational Intelligence, the ACM, IEEE, GI, the Austrian Computer Science and the Association for the Advancement of Artificial Intelligence (AAAI). Since 2003 Andreas has participated in leading positions in 30+ R&D multi-national projects, budget 6+ MEUR, 300+ publications, 11k+ citations, h-Index = 47.


Abstract
In the medical domain the expectations to automatic AI/machine learning systems are extremely high, particularly in disciplines requiring prognostic models (oncology) and/or decision support (radiology, pathology). Due to raising ethical, social, and legal issues governed by the European Union, the problem of replicability, understandability, explainability and transparency is becoming important. The progress in probabilistic machine learning, the availability of large amounts of training data and increasing computational power has made AI/machine learning successful today, and in certain medical tasks it even exceeds human performance. However, such approaches are considered as “black box”- models, and even if we understand the underlying mathematical principles of such models, they still lack explicit declarative knowledge. Most of all these models are very sensitive to even slight perturbations. Consequently, the goal in health informatics is to make machine learning models more robust. One possible step is in linking probabilistic learning methods with large knowledge representations, thus allowing a human expert to understand how a machine decision has been reached and to interact on demand. Our aim is not only to make machine decisions re-traceable, interpretable and comprehensible, but to interpret why a certain machine decision has been reached. In medicine the “why” is often more important than the classification result. The re-traceability and interpretability on demand shall foster reliability and trust ensuring that the human remains in control, so to augment human intelligence with artificial intelligence and vice versa.



footer