- AMDS@Pacmed #1: Bayesian Deep Learning & Rheumatoid Arthritis
Organised by Pacmed in collaboration with Freedomlab. Programme: 18.00 - 18.30 Doors open 18.30 - 19.00 David Ruhe 19.00 - 19.30 O'Jay Medina 19.30 - 20.30 Drinks There will be some (small) snacks and drinks before and after the event. www.pacmed.ai www.freedomlab.org First speaker: David Ruhe (Pacmed, MSc Data Science) Bayesian Modelling in Practice: Using Uncertainty to Improve Trustworthiness in Medical Applications. This talk will illustrate how the predictive uncertainty that Bayesian models provide can be used to mitigate risk of misguided prediction and to detect out-of-domain examples in a medical setting. Using the MIMIC-III critical care dataset, a Bayesian Neural Network was trained to predict mortality after discharge from the Intensive Care Unit. Empirical results show that such approach can indeed be used in practice to prevent potential errors and reliably identify observations for which the prediction is uncertain. David Ruhe researches methods to assess uncertainty of machine learning predictions. Specifically, he is looking at modeling uncertainty with Bayesian Neural Networks applied in Pacmed's Intensive Care Unit clinical decision support system. Second speaker: O’Jay Medina (UMC Utrecht) Rheumatoid arthritis (RA) is an autoimmune disease causing inflammations in joints during periods of high disease activity (flares). No permanent cure is available and treating with biologicals is both expensive and affecting the quality of life in patients. Empirical evidence shows that flares can be predicted based on patient data available in electronic health record systems by using data science and machine learning techniques. Using this evidence, a dynamic prediction tool is being developed to improve clinical decision-making in RA treatment. O’Jay Medina is a Data Scientist at the University Medical Center Utrecht (UMCU). He has a background in neuroscience and is currently part of the Applied Data Analytics in Medicine (ADAM) program. Apart from developing software to predict Rheumatoid Arthritis flares, he is a member of the UrStatus project, which aims to provide an intuitive visualisation of the status of patients in the Neonatal Intensive Care Unit (NICU) by using predictive modelling.
- 11th Medical Data plus Pizza meeting
ONE THOUSAND MEMBERS AND COUNTING (!) Everything medical data science. And pizza. This edition features: ** Deep learning for Tumor Response Evaluation Geert Kazemier (Surgery, Amsterdam UMC) ** Why you should use our Research Infrastructure Arno Sinjewel (ICT, Amsterdam UMC) followed by discussion and networking with pizza
- Amsterdam UMC Bioinformatics Seminar featuring Esther Smit
Dear Colleagues, You are cordially invited to the 2nd symposium of the Amsterdam UMC bioinformatics community: Amsterdam UMC Bioinformatics Seminar 3rd of June 2019 - 17:00-18:00 AI ecosystems and public-private partnerships Guest Speaker: Dr. Esther Smit, Partnerships Manager at Innovation Center for Artificial Intelligence Location: Fonteinzaal - C0 (Amsterdam UMC, location AMC) Within Amsterdam there are strong ecosystems around AI and Data Science. Examples include Amsterdam Data Science and of course Amsterdam Medical Data Science, where academics and industry are partners looking into possibilities to work together and support the Amsterdam ecosystem. Another example is the Innovation Center of Artificial Intelligence (ICAI), a national initiative which started in Amsterdam. Here, researchers in AI work with industry partners on research agenda’s with a strong external outreach for a period of 5 years. Within this talk, Esther will give an overview of the different ecosystems, how they are organised and how you can be part of them. IMPORTANT: For those of you who haven’t yet filled out our survey please do so here: https://forms.gle/pJE3pHoE2XnvKv9P8 This information will help us identify your needs as a community and to make a strong case to get those needs met by the ICT infrastructure… We will present the outcome of the survey prior to the lecture of Esther on June 3.
- AMDS special: Integrative Learning for Heterogeneous Data from Clinical Studies
Auditorium – O|2 Building
Organized by Viktor Wottschel in collaboration with EuroPOND - www.europond.eu EPAD - www.ep-ad.org and AMYPAD - www.amypad.eu Modeling the relationship among heterogeneous data is essential in clinical studies, where understanding the dynamics of neurodegeneration requires to jointly account for the heterogeneity of demographic, clinical, and imaging information, along with confounders from biology and genetics. While standard approaches based on machine learning are finding increasing popularity, their clinical applications if often frustrated by the lack of interpretability and robustness of the findings. The aim of this presentation is to illustrate our contribution towards interpretable and robust learning methods for the analysis of data from clinical studies. In particular, we will focus on approaches for the automatic identification of the joint relationship among high-dimensional biomarkers, such as imaging, genetics and neuropsychological tests. Moreover, we will illustrate our advances in the modelling of the long-term natural history of the pathology from short-term longitudinal observations from clinical trials. We will show how our methods can automatically identify plausible progressions of several disease progression biomarkers, including full resolution medical-images from different modalities. We will finally show how our methods can be used as statistical reference for automatic staging and prediction in unseen clinical trial data. Marco is a group lead at INRIA Sophia Antipolis with research interests in Statistical Learning, Multimodal Brain Image Analysis, Imaging Genetics, Computational Clinical Neuroscience and Spatio-temporal Models. Prior to this, he obtained his PhD from the University of Nice in 2012, worked as a statistical consultant for the Pharmacog project, and was a senior research associate at INRIA and UCL.
- 10th Medical Data plus Pizza meeting - Pitch Edition
Amsterdam UMC - Location VUmc - Lecture Hall Rijn
--- Program --- *** Hosted by Lucas Fleuren (Intensive Care, Amsterdam UMC) *** Welcome Dr. Paul Elbers (Intensive Care, Amsterdam UMC) Dr. Mark Hoogendoorn (Computational Intelligence VU) *** Opening Prof. dr. Guus Schreiber, dean Faculty of Science (VU) *** Pitches: - Michael Kemna (TU Delft) - The Netflix Experience for Hospitals - Edwin Geleijn (Amsterdam UMC) - Text Mining to Prevent Falls - Brian Doelkahar - (ABN AMRO) Anomaly Detection Model within Health Care - Helen Gosselt - (Amsterdam UMC) Therapeutic Drug Monitoring of MTX in Rheumatoid Arthritis - Annanina Koster (Ynformed) - Predicting PROMS for Shared Decision Making *** Intermezzo Prof. dr. Armand Girbes, head Intensive Care (Amsterdam UMC) *** Key Note Lecture: aI Amsterdam Jeroen Maas, Amsterdam Economic Board *** Pitches: - Patrick Thoral (Amsterdam UMC) - Save a Life: Predicting Serious Adverse Events - Martijn Schut (Amsterdam UMC) - Mining with Meaning - Patrycja Książek (Amsterdam UMC) - Pupillometry to Detect Listening Effort - Tom van den Bosch (Amsterdam UMC) - Predicting Complications after Cancer Surgery - Melvyn Roerdink (Amsterdam UMC) - HoloLens Mixed Reality for Alleviating Freezing of Gait in Parkinson's Disease *** Pizza plus networking --- Event info --- Find a doctor or data scientist Are you a doctor or other healthcare professional? And do you want to learn more on big data and machine learning? Or do you have a great idea for a machine learning application that could benefit from a data scientist? Are you a data scientist and do you want to work with doctors and other healthcare professionals to work on projects that matter? Or do you want to meet them to discuss and improve your own projects? This is you chance to listen, meet or pitch! Amsterdam Medical Data Science unites 700 healthcare professionals and data scientists. With an important mission: to improve patient outcomes using big data and machine learning. Join us for a special event! What? Find a Doctor or Data Scientist. Who should attend? Everyone interested in machine learning and big data, especially doctors, other healthcare professionals and data scientists. Goal? Meet and Collaborate. Format? Ten pitches, four minutes each. When? Tuesday, May 21 at 1700 hours. Where? Amsterdam UMC, locatie VUmc, Rijnzaal. Food? Pizza after the pitches Registration? www.amsterdammedicaldatascience.nl (use chrome) Want to pitch? [masked] or via meetup
- AMDS special: Bioinformatics Seminar with Guszti Eiben
Amsterdam Medical Data Science is proud to announce a special meetup: the first Amsterdam UMC Bioinformatics Seminar has invited Prof. Dr. Guszti Eiben (VU) as guest speaker. Prof. Dr. Eiben is specialized in the field of artificial intelligence, Artificial Life, and Adaptive Collective Systems. and he leads the Computational Intelligence Group. His approach to AI is based on Evolutionary Computing and over the last 30+ years he worked on a diverse set of topics including: theoretical foundations and applications in health, finance, and traffic management. Prof. Eiben also built a system to evolve Mondriaan and Escher style art and exhibited it in the Haags Gemeentemuseum. His current research is focused on evolutionary robotics. For more impressions on Prof. Dr. Eiben, watch his interview on Nieuwsuur Links: https://www.cs.vu.nl/ci/ http://www.focas.eu/adaptive-collective-systems/ https://www.springer.com/cn/book/9783662448731 https://nos.nl/uitzending/39884-nieuwsuur.html Hosted by Bart Westerman (Amsterdam UMC, location VUmc) and Elisabeth Lodder (Amsterdam UMC, location AMC)
- AMDS special: Deep Learning to Build an Atlas of the Human Brain
Organized by Team Clinical Neuroanatomy and Biobanking FreeSurfer enables morphometric, functional, and connectivity studies of the human brain in vivo using MRI. While FreeSurfer is frequently updated to stay near the state of the art in terms of methodology, it still relies on a computational atlas derived from in vivo MRI scans acquired over a decade ago, and which fail to describe the human brain beyond the whole structure level. In this talk, I will present ongoing work to build a computational atlas of the whole human brain at the substructure level, which we intend to integrate with FreeSurfer. First, I will present our initial work on an atlas of the hippocampal subfields using ex vivo MRI, which represented the first step towards building such an atlas of the whole brain. Both the construction of the atlas and its application to automated image segmentation rely on Bayesian inference techniques within a generative framework of imaging data. Next, I will explain the limitations of ex vivo MRI in atlas building, and how we are overcoming them with histology. Unfortunately, histological analysis introduces a new set of problems associated with the geometric distortion caused by sectioning and processing; I will present techniques based on Bayesian modeling and weakly supervised deep learning that we have developed to correct for this distortion. Finally, I will present preliminary results on population studies of the hippocampus, thalamus and amygdala, as well on the 3D reconstruction of the publicly available atlas of the Allen Institute. Juan Eugenio Iglesias holds two MSc degrees in telecom and electrical engineering from the University of Seville (Spain) and the Royal Institute of Technology (KTH, Sweden), respectively. After two research assistantships at Seville and the University of Copenhagen (Denmark), he carried out his PhD in Biomedical Engineering at the University of California, Los Angeles (UCLA) with a Fulbright grant. He was subsequently a postdoctoral research fellow at the Martinos Center for Biomedical Imaging and at the Basque Center on Cognition, Brain and Language (Spain), the latter funded by a Marie Curie fellowship. In September 2016 he joined University College London (UK) as a lecturer with a Starting Grant of the European Research Council. Since November 2019, he is visiting MIT, and working part time at the Martinos Center.
- 9th Medical Data plus Pizza meeting
Everything medical data science. And pizza. This edition features Ari Ercole, University of Cambridge, UK Machine Learning for Intensive Care Medicine Ronald Cornet, Amsterdam UMC, Location AMC From care data to FAIR data followed by discussion and networking with pizza
- 8th Medical Data plus Pizza meeting
Everything medical data science. And pizza. This edition features Sanne Abeln - VU University Amsterdam Bioinformatics for biomedicine: looking for a needle in a haystack? Kai Lønning - Spinoza Institute for Neuroimaging Deep Learning for MRI followed by discussion and networking with pizza
- 7th Medical Data plus Pizza meeting - Special Edition (!)
Amsterdam UMC - location VUmc - ICU - Room Delta
Everything medical data science. And pizza. This edition features the first author of the recent Nature Medicine paper on machine learning in sepsis. You should attend! Mathieu Komorowski - Imperial College, London - UK Reinforcement learning to optimise sepsis treatment followed by discussion and networking with pizza