2021 - AI4Health

2021 - Artificial Intelligence AI4Health

The Lecture Series continues on 27 January 2021 as WebEx Online Meetings

............................... WEBEX Access ...............................

We invite you to join this Webinars.

Meeting link:(but please check the invitation as the link may change)
https://unilu.webex.com/unilu/j.php?MTID=m321de1741e5670737950863dd1982d5a

Meeting number (access code): 175 952 5790
Meeting password: SujKPJ5MB52 Host key: 912098

........................................................................

Thanks to the CLAIRE network!

....................................... Contact .......................................

• Dr. Vladimir Despotovic (Info), Department of Computer Science
• Prof Dr Jorge Goncalves (Info), LCSB
• Dr Jun Pang (Info), Department of Computer Science
• Dr. Venkata Satagopam (Info), LCSB
• Prof Dr Thomas Sauter (Info), Department of Life Sciences and Medicine
• Prof Dr Christoph Schommer (Info), Department of Computer Science

...................................... Schedule ......................................

Speakers

Prof Dr Rafael Orozco-Arroyave, University of Antioquia + Adjunct researcher at the Pattern Recognition Lab, University of Erlangen

An overview of ML methods to model symptoms in movement disorders (the case of Parkinson’s disease): from classical ML to DL

There exist different movement disorders with different origin. Parkinson’s disease (PD) is one of those disorders and appears due to the progressive death of dopaminergic neurons in the substantia nigra of the mid-brain. Diagnosis and monitoring of PD patients is still highly subjective, time consuming and expensive. Existing medical scales used to evaluate the neurological state of PD patients cover many different aspects, including activities of daily living, motor aspects, speech and depression. This makes the task of automatically reproducing experts’ evaluation very difficult because several bio-signals and modeling methods are required to produce clinically acceptable/practical results. This talk tries to show part of the way that has been traveled since about ten years considering different bio-signals (e.g., speech, gait, and handwriting) and methods of Machine Learning and the relatively new topics of Deep Learning (DL) with the aim to find suitable models for PD diagnosis and monitoring. Results with classical feature extraction and classification methods will be presented and also experiments with CNN and LSTM architectures will be discussed.

Biography: Juan Rafael Orozco-Arroyave was born in Medellín, Colombia in 1981. He is an Electronics Engineer from the University of Antioquia (2004). From 2004 to 2009 he was working for a telco company in Medellín, Colombia. In 2011 he finished the MSc. degree in Telecommunications from the Universidad de Antioquia. In 2015 he finished the PhD in Computer Science in a double degree program between the University of Erlangen (Germany) and the University of Antioquia (Colombia). Currently Juan Rafael Orozco-Arroyave is an Associate Professor at the University of Antioquia and adjunct researcher at the Pattern Recognition Lab at the University of Erlangen.

Prof Dr Sergio Martinez-Cuestra

AstraZeneca R&D and the University of Cambridge

Developing new genomics technologies to map DNA epigenetic modifications and damage in humans, parasites and cancer

Abstract
DNA is much more than ACGT. I will give a broad overview of the experimental and computational technologies and challenges used to map and understand epigenetic modifications, damage and structures in DNA. We are just beginning to understand how changes to the chemistry of DNA play a key role in the development of diseases (e.g. cancer), life cycle control and development of eukaryotic diseases. Using recent projects from my own collaborative research with talented chemists and biologists I will introduce the state of the art in the areas of DNA epigenetics from a computational perspective. 

Biography

Sergio is starting a research group in bioinformatics exploring fundamental mechanisms of protein degradation and DNA epigenetics at the interface between industry and academia. He builds from his early training in laboratory chemistry and biochemistry, through a PhD and postdoc appointments in bioinformatics and cheminformatics developing projects in collaboration with wet-lab colleagues sharing his time between the AstraZeneca headquarters and the University of Cambridge. Outside research, Sergio teaches bioinformatics at university, contributes as an editor of the emerging journal Frontiers in Bioinformatics and supports the growth of the European life science data infrastructure ELIXIR. Beyond science, he has a passion for team sports, languages and culture/gastronomy. 

• https://www.linkedin.com/in/sergio-mart%C3%ADnez-cuesta-a2525437/
• https://scholar.google.com/citations?user=qRdu6_YAAAAJ

Dr Mauro Dragoni, Fondazione Bruno Kessler, Trento, Italy

Achieving Explainable AI Through Semantic Technologies: Challenges and Future Directions in Digital Health 

Abstract The interest in Explainable Artificial Intelligence (XAI) research area has dramatically grown during the last few years. The main reason is the need of having systems that beyond being effective are also able to describe how a certain outcome has been reached and to present it in a comprehensive manner with respect to the target users. The Digital Health domain is a pioneering scenario where XAI strategies have been designed and implemented. In this talk, I would go through the milestones of this paradigm and I will discuss the role of semantic technologies. Then, I will present how such strategies have been applied to the Digital Health domain, and which are the challenges that have to be tackled in the near future.

Short Bio Mauro Dragoni is a research scientist at Fondazione Bruno Kessler within the Process and Data Intelligence (PDI) Research Unit. He received his Ph.D. in Computer Science from the University of Milan in 2010. His main research topics concern knowledge management, information retrieval, and machine learning with a focus on the design and development of real-world prototypes for enabling the access of both industries and people to research. Since 2015, he has been involved in activities dedicated to bring AI solutions within the Digital Health area. He has been involved in a number of national and international research projects and he co-authored more than 100 scientific publications in international journals, conferences, and workshops. Beyond AI, he is a motorsport (real and simulated) passionate and a certified personal trainer from the Italian National Olympic Committee.

Dr Aureli Sonia-Frisch, Director, Neuroscience BU, Starlab Barcelona

Machine Learning for Brain Health and Understanding at Starlab Neuroscience

Abstract

The convergent development of different technologies is bringing the understanding of the brain, both on healthy and pathological condition, further than ever before. The confluence of Wearables, Neurotechnologies, Augmented and Virtual Reality, Serious Gaming, Data Science, Machine Learning and Artificial Intelligence are a game changer in the way we study brain functionality, make use of it for interacting with the environment, and treat mental and neurological disease. The talk will deal with the combination of Neurotechnologies, Machine Learning and Artificial Intelligence in different Digital Brain Health applications developed at Starlab Neuroscience. Digital markers of brain function will lead in the near future to improved diagnostic, drug discovery, risk analysis, and interactivity. We will show developed methodologies for: stratified performance evaluation of classifiers in operational conditions for Parkinsons’ risk assessment, differential diagnosis in ADHD based on Reservoir Computing, and new treatment outcome prediction in Coma patients. I will go over the technical challenges we faced to develop these applications, but also over some insights that influence the applicability of pure academic data science in the real world.

Biography

Dr.-Ing. Aureli Soria-Frisch received the MSc from the Polytechnic University of Catalonia– UPC (1995), and the PhD from the Technical University Berlin. Between 1996 and 2005 he worked at the Fraunhofer IPK (Berlin), as research scientist and project leader. After 3 years as Visiting Professor at the Universitat Pompeu Fabra, he joined Starlab in 2008. He is the Director of the Neuroscience Business Unit since beginning 2017. His research interest is focused on the fields: computational intelligence for data analysis, and machine learning for electrophysiological signal analysis. He has authored 20 journal papers, seven book chapters, and over 60 conference papers. He has been Project Manager of the FP7 HIVE project, where the Starstim early prototype for transcranial electrical stimulation was developed, Project Coordinator of the H2020 FET Open LUMINOUS project on the clinical study of consciousness, and PI of the 2 MJFF grants for the development of Machine Learning PD biomarkers. 

RUDI BALLING GOODBYE LECTURE SERIES (Talks 1 - 5)

see also the annoncement of the Luxembourg Centre for Systems Biomedicine HERE

Talk 1 :
Prof. Rudi Balling: Wicked problems: Mission impossible or the next frontier?

Abstract:
Today’s problems cannot  be solved any more by a single person, organisation or discipline. Therefore, interdisciplinary cooperation and systems approaches have been widely adopted to cope with the complexity and uncertainty of our world. Big data and artificial intelligence penetrate almost every single aspect of our lives. However, we now realise the limitations of such a data driven toolbox. There are problems that seem to defy a solution. These are sometimes called “wicked problems”, that apparently have many possible, but no real solution. Each wicked problem is unique and involves social or cultural issues, touching upon a diversity of individual or societal values. As a result, a much deeper understanding of the stakeholders involved is necessary. I will give an overview and discuss some of the challenges related to wicked problems.

Talk 2 :
Prof. David Leigh: Making the tinest machines

Abstract: Perhaps the best way to appreciate the technological potential of controlled molecular-level motion is to recognise that molecular machines lie at the heart of every biological process. Nature has not repeatedly chosen this solution for achieving complex task performance without good reason. In stark contrast to biology, none of mankind’s myriad of present day technologies exploit controlled molecular-level motion in any way at all: every catalyst, every material, every pharmaceutical, all function through their static or equilibrium dynamic properties. When we learn how to build artificial structures that can exploit molecular level motion, and interface their effects directly with other molecules and the outside world, it will potentially impact on every aspect of functional molecule and materials design. An improved understanding of physics and biology will surely follow. 

Talk 3:

Prof. Ortwin Renn: Behavioral adaptations to the COVID-19 crisis: What is here to stay?

Abstract: The rise of populism and corresponding political resonance in Europe and the USA is precarious because it is threatening democracy and science. Many political decision-making processes are based on evidence and expert knowledge. This is particularly true for systemic risks such as the recent COVD-19 pandemic. However, many political and social actors discredit wellgrounded knowledge as “fake news” or conspiracy theory; they bring alternative facts and truth into play. The public is often confused and lacks orientation. The paper will stress the importance of robust knowledge stemming from science, expertise and practical experience.

Talk 4:

Dr. Sara-Jane Dunn: Biological Computation in Stem Cells

Abstract: Experimental biology has proven our ability to induce cell identity via differentiation or reprogramming, offering huge promise for medicine and the study of development. Yet despite this wealth of research, an explanation of how cell state conversions arise remains fragmentary. Ideally, we would like to understand the complex, dynamic interplay of genetic components that manifests as cell fate conversions. To address this gap, computational analyses can be combined with mathematical modelling to interrogate experimental data and generate testable hypotheses on how a program of genetic interactions governs cell identity. In this talk, I will demonstrate how interdisciplinary approaches have revealed the biological program governing fate decisions in stem cells, and indeed, could be used in other domains to expose the regulatory programs that drive cellular decision-making more broadly.

Talk 5:

Prof. Edith Heard: Launching a new area of biology to unterstand 'Life in context'

Abstract: coming soon

Dr. habil. Jürgen Landes: Causal Inference in Medicine in the real world

Abstract: In an ideal world, we draw causal inferences based on well-controlled randomised experiments carried out by independent and impartial experts. In reality, we cannot infer whether drugs cause adverse reactions from such data. Randomised clinical trials are too short and study too few patients to observe rare (and possibly severe) adverse reactions. These trials are often carried out by employees of a multi-billion dollar industry incentivised to sell their products. In this talk, I will i) present a Bayesian framework, E-Synthesis, which > facilitates causal inference from (possibly biased) real world evidence and ii) discuss its applicability in the real world.

Bio: After obtaining a PhD in mathematical (probabilistic) logic and two brief postdocs I turned to philosophy in 2012. I have since been interested in evidence and confirmation of (causal) hypotheses. In particular, I'm much interested in how to assess the causal hypothesis that a drug causes an adverse reaction based on real world evidence. I also work on Maximum Entropy inference (finite and infinite domains) and Bayesian epistemology of science.

Prof. Dr. med. Joachim Klucken: Integration of AI into the life of patients and healtcare providers: how does it work and what does it change?

Abstract:
Wearable sensors and smartphone apps are increasingly providing information - usually referred to as "data" - from the real world environment of patients home. Data-driven medicine has generated a profound area of research aiming to provide better diagnostic and therapeutic applications. The anticipated effects for patients as well as the market for new digital healthcare services seems endless. Using the example of wearable sensors on gait analysis in Parkinson's disease the presentation will show how to translate data-driven research into data-driven medicine. The technical, medical and societal aspects of the different development phases of data-driven medicine will be presented, as well as the difference between AI-driven innovations and AI-driven medical applications. The goal is to better understand how to bring smart algorithms and real-world data driven innovations into healthcare applications that in the end are beneficiary for patients, healthcare provider and society.

Vita:
Digital medicine is a new field Medicine that aims to understand how patient-centered technology can be used in everyday medical practice, and which evidence assessment is needed to not only understand the medical benefits of healthcare technologies, but also their patient- and social acceptance and economical efficacy. Here, the major goal lies in clinical studies for healthcare technologies providing evidence for their medical, social, ethical and legal benefit as well as economic efficiency ultimately generating a concept of “clinical validation of healthcare technologies and services”. Prof. Klucken earned his MD in Laboratory Medicine and specialized in Neurology. He finished his habilitation thesis in 2009 in translational neuroscience in Parkinson’s disease including work at the Massachusetts Institute for Neurodegenerative Diseases, Harvard Medical School, Boston, USA on neurodegenerative processes in Parkinson’s disease. In 2004 he also started translational research projects in the field of medical technology (m/eHealth) applying sensor-based motion detection in movement disorders. Jointly with engineers and data-scientists, he developed novel gait-specific instrumented movement analysis concepts for Parkinson’s disease, multiple sclerosis, osteoarthritis, sarcopenia, oncology and healthy well-being of the elderly. From 2008 until 2021 he was a senior physician and PI at the Movement Disorder Unit (Department of Molecular Neurology, University Hospital Erlangen, Germany) and developed sensor-based gait analysis for patients with movement disorders. From 2018-2021 he also lead a group at Fraunhofer IIS, Erlangen, Germany with the focus on developing digital health pathways that enable technology integration into healthcare workflows. In 2019 he also established a contract research organization (Medical Valley Digital Health Application Center - dmac) supporting personalized healthcare technologies in order to get access to the German healthcare market. Within the scientific community he initiated and leads the task-force “Telehealth Services” of the Germany Parkinson Society (DPG), he is a founding member of the task-force “technology” of the international movement disorder society (MDS), and he is the chairman of the advisory board “e-health, telematics methods) of the Professional Association of German Neurologists (BDN). On political and societal level including patient-support groups he promotes the use of mobile healthcare technologies and innovations for comprehensive digital healthcare services, clinical studies and care. In addition, he participates in spin-offs/start-ups in the field of sensor-based movement analysis, and is advising several pharmaceutical companies and healthcare insurances/services on the topic of wearable derived objective outcomes.