Awards Take-off phase 1 - cluster Commit2Data
RSS Feed
This article was placed on: c2d
The Take-off call is intended for entrepreneurial scientists who want to investigate whether the results of their research have commercial potential. In these calls, which are aimed at stimulating business activity, research projects can apply for money to do market research, among other things. Commit2Data has its cluster within the Take-Off calls and the following projects have been awarded a grant from it:
Low-Code Development Platform for Security of AI
DISTANT - Dr. Stjepan Picek – Radboud University
Machine learning models achieved excellent performance for various real-world applications due to their practicality and effectiveness. Unfortunately, the widespread use of machine learning also opens new security threats due to various failure modes of machine learning. The evolution of secure and robust AI systems depends substantially on understanding new threats and failure modes. This project explores the viability of a start-up company providing a framework that allows evaluating the security of AI against various threats. The framework provides a low-code development platform for security and machine learning practitioners, collaborative work options, and state-of-the-art functionalities.
A platform to easily and safely apply artificial intelligence for the electrocardiogram in clinical care
Dr. René van Es – University Medical Center Utrecht
The electrocardiogram (ECG) was introduced over 100 years ago, but it is still interpreted by physicians in the same way. Recently, artificial intelligence (AI) algorithms have shown to be able to interpret the ECG faster and
more accurately. Moreover, AI can detect abnormalities in the ECG, even before they become apparent to the physician. Despite this, no algorithms have been implemented in the daily clinical workflow, as the current software platforms do not allow this. We will develop the vendor-neutral ECG platform of today, that allows for advanced ECG analysis and easy and safe implementation of ECG-AI algorithms.
RAILWAY+: Feasibility of a Personal Health Train Service for Complex Health Data Partitions
Prof. Dr. Andre Dekker – Maastricht University
Unleashing the potential of real world data is seen as key to developing and introducing innovations in health care, due to its high representativity, volume, and low cost. Alas, these data remain scattered, protected in silos and out-of-reach for researchers. The Personal Health Train (PHT) concept promises to enable researchers to access the data, but its implementation requires a complex combination of specialised software and legal agreements and is currently only available for horizontally partitioned data. Our aim is to offer a comprehensive solution to implement the PHT as a service for horizontally and vertically partitioned data.