Negen interdisciplinaire AI-projecten worden door de alliantie gefinancierd met als doel interdisciplinaire samenwerking tussen de vier alliantiepartners en maatschappelijke partners aan te moedigen, om zo de ontwikkeling en toepassing van AI ten voordele van de samenleving te stimuleren.
De geselecteerde projecten bestrijken een scala aan innovatieve AI-methoden die bijdragen aan de overgang van reactieve naar meer proactieve en gepersonaliseerde zorg, waaronder zorg op afstand en zelfzorg, maar ook aan primaire preventie en het terugdringen van administratieve lasten. Naast de toepassing van AI in preventie en in de gezondheidszorg, worden methoden ontwikkeld om AI betrouwbaarder, valide en transparanter te maken. AI for Health werkt samen met het Institute for Preventive Health van de alliantie, dat een van de geselecteerde projecten financiert.
Neem contact op met ons AI for Health team als je meer wilt weten over de projecten of over het thema AI for Health van de alliantie.
De bekroonde projecten zijn:
DIgitally Nudging Outdoor play, Game AI for MEntal Health
DINO GAME aims for the development of AI in a multiplayer location-based game to optimize individual physical activity, particularly for children who have limited school attendance and play with peers due to chronic illnesses or Covid-19. While UU and UMC Utrecht already collaborate on games and gamification for children with a chronic illness, the focus in DINO GAME is on the physical play. The collaboration with WUR will bring in complementary expertise on the area of spatial cognition and applied gaming. The project reinforces collaboration with universities of applied sciences (HU and HvA), and societal partner, the Trimbos Institute. The project is funced by the alliance’ Institute for Preventive Health.
Research team: Alexander Klippel (WUR), Sander Bakkes, Heidi Lesscher, Remco Veltkamp (UU – Lead), Kors van der Ent, Sanne Nijhof (UMC Utrecht), Lisette van der Poel (HU), Marloes Kleinjan (Trimbos Instituut), Annette Brons (HvA, Digital Life)
A patient-centric lifestyle recommender system for optimizing treatment response in psychosis
U-HEAL aims at finding explainable AI solutions to improve the personalized treatment response in patients with psychosis. The project focuses on the development of a human-centered and explainable AI framework for a smart healthcare application in psychiatry that promotes distant self-care. The project combines expertise on data analysis with a focus on machine learning in psychiatry (UU), machine learning solutions (UMC Utrecht), uncertainty in AI, recommendation systems, and counterfactual learning for model explainability (TU/e), and nutrition and medicines (WUR).
Research team: Maryam Tavakol (TU/e), Renger Witkamp (WUR), Hugo Schnack (UU – Lead), Wiepke Cahn, Seyed Mostafa Kia (UMC Utrecht).
Generating pilot data through AI for deep-phenotyping of atherosclerotic plaques and the discovery of novel biomarkers for patient stratification
PlaqAI aims to improve and refine the phenotyping of atherosclerotic lesions through a data driven and agnostic approach. The project has the ambition to achieve higher success rates in disease-associated biomarker discovery which could be used for preventive healthcare, or in the setting of drug targeting as surrogate markers of efficacy. The project integrates different ‘omics’-datasets from the Athero-Express Biobank Study with the knowledge and expertise on atherosclerotic genetics (UMC Utrecht), image-based analyses (TU/e) and AI (UU).
Research team: Mitko Veta (TU/e), Alejandro Lopez Rincon (UU), Sander van der Laan (UMC Utrecht – Lead)
X plAIn-me Please
Explainable AI and Patient Participation in Prediction and Prevention of aneurysm rupture
X plAIn-me aims to leverage AI for the development of personalized prediction models of aneurysm rupture risk, acquisition of knowledge on modifiable lifestyle risk factors for aneurysm rupture and improvement of patient participation and health care professionals education. The project focuses on the implementation of new technical developments through medical education in the clinic setting (TU/e, UU, UMC Utrecht) and the use of machine learning techniques to investigate diet and nutrition risk factors for aneurysm rupture from the big dataset of the UK biobank (TU/e, WUR, UMC Utrecht).
Research team: Remco Duits (TU/e), Marianne Geleijnse (WUR), Marieke van der Schaaf (UU), Irene van der Schaaf, Ynte Ruigrok, Hugo Kuijf, Birgitta Velthuis (UMC Utrecht – Lead)
Automated medical reporting to reduce the administrative burden on healthcare providers: Configuration of the linguistic pipelines
Care2Report aims to reduce the amount of time spent on administration tasks among healthcare professionals using speech recognition and action recognition technology to automatically record and summarize an interaction between a care provider and a patient. The project uses the complementary expertise on product software development and open-source software platforms (UU), digitisation projects involving outpatient care (UMCU), and use of graphical architecture tools for linguistic AI (TU/e).
Research team: Michel Chaudron (TU/e), Sjaak Brinkkemper, Fernando Castor de Lima Filho (UU – Lead), Marije Marsman, Annemarie van ‘t Veen, O’Jay Medina (UMC Utrecht)
Better Imputation by Generative Adversarial NeTworks
BIGANT aims to smarten Multivariate Imputation by Chained Equations (MICE) by adopting recent advances in machine learning and AI to address the issue of incomplete data sets. The project uses the complementary expertise on solving missing data problems by MICE (UU), a deep understanding of and experience with applying Generative adversarial networks across different fields (TU/e), solving missing data problems in epidemiological and clinical application and developing novel data techniques for real-time monitoring and predicting health (UMCU), and modelling of longitudinal patterns in plant and animal systems as functions of genetic, environmental and time (WUR).
Research team: Mykola Pechenizkiy, Cassio de Campos, Rianne Schouten (TU/e), Fred van Eeuwijk, Hendriek Boshuizen (WUR), Stef van Buuren, Gerko Vink, Hanne Oberman (UU – Lead ), Daniel Oberski, Thomas Debray (UMC Utrecht)
Early detection of burnout risk in healthcare workers
Integrating psychological and physiological data in an AI-based framework
The project investigates the potential for an AI-based early warning system for burnout by analyzing both psychological and physiological data with conventional and machine learning techniques. The early detection system for burnout is the stepping stone towards timely strategies to prevent burnout development, thereby contributing to creating and maintaining a healthy and resilient workforce. The project uses the complementary expertise on psychological work-related factors that are related to burnout (UU), physiological person-related factors associated with burnout (TU/e), AI techniques and its application in the health domain (WUR), and developing, validating, evaluating, and implementing an AI Prediction Algorithm in the medical sector (UMC Utrecht).
Research team: Leander van der Meij (TU/e), Hubert Fonteijn (WUR), Maria Peeters (UU – Lead), Jaap Trappenburg (UMC Utrecht)
The scientific foundation for an online self-help platform for dealing with loss and grief
The project proposes an investigation into the creation of an AI-driven platform for grief support at home. The platform will support users in finding the most effective coping style that fits their current needs, which requires a supervised machine learning approach to cater to personalisation. The project uses the complementary expertise in AI and human computer interaction, ethics of socially disruptive technology (TU/e), cognitive behavioural therapy, applied data scientist (UU), psychiatry, psychopathology (UMC Utrecht), and data visualisation, interactive multimedia systems and software engineering (WUR).
Research team: Sanne Schoenmakers, Wijnand IJsselsteijn, Rooks (TU/e – Lead), Wil Hurst, Kwabena Ebo Bennin (WUR), Paul Boelen, Matthieu Brinkhuis (UU), Manik Djelantik (UMC Utrecht)
Improving the health of colorectal cancer patients through developing and applying causal Artificial Intelligence
The project aims to identify lifestyle factors that are most important in the course of disease of colorectal cancer (CRC) patients, which is a crucial first step to reach the ultimate goal of establishing evidence-based dietary and lifestyle guidelines to improve the health of CRC patients. The project proposes to combine existing AI approaches geared towards predictive modelling with new and existing methods for causal discovery. The project uses complementary expertise on data science (UU), electrical engineering, causal modelling in AI (TU/e), cancer epidemiology, clinical epidemiology and medicine (UMC Utrecht), and nutrition, and physical activity (WUR).
Research team: Fons van der Sommen (TU/e),Fränzel van Duijnhoven (WUR – Lead), Oisin Ryan (UU), Anne May, Miriam Koopman (UMC Utrecht)