PhD researcher in explainable machine learning and survival analysis with an application to smart hospital alarms
itec is a research group of KU Leuven and imec, Flanders’ high-tech research and innovation hub for nanoelectronics and digital technologies. The group conducts interdisciplinary research on the design, development, and evaluation of personalized and adaptive digital solutions, with applications mostly in the domain of technology-enhanced education and training, but also in media and in health. In order to realize this, itec brings together researchers trained in the learning sciences, statistics, computer science, and applied linguistics, in a cooperative research lab on the Kulak campus in Kortrijk. The research group often collaborates with industry and societal partners.
The main focus of the AI and machine learning subgroup is to apply existing and develop new machine learning algorithms to advance the application domains. In particular, we focus on supervised, unsupervised and semi-supervised learning. Key activities in the group relate to multi-label/target prediction, interaction prediction, recommender systems, active learning, survival analysis and longitudinal data analysis; always with a specific attention to interpretability.
Project
This PhD position,which is funded by the Research Foundation Flanders (FWO), consists of two main parts. First, the candidate will carry out fundamental research on the intersection between interaction learning and survival analysis. In particular, interaction learning – also called pairwise learning or network mining – is situated in the machine learning domain. It deals with predicting or clustering interactions between two sets of objects (e.g. users and products). Survival analysis is rooted in statistics and deals with predicting the time until an event occurs. Recently survival analysis has entered the machine learning field. The research group has acquired expertise in both domains, however the combination of interaction learning and survival analysis is novel. The candidate will develop new algorithms for prediction and clustering in this context. Second, the developed algorithms will be applied in the context of developing smart alarms in an intensive care unit. Given the current overload of alarms, the candidate will contribute to reduce alarm fatigue in clinical staff and alarm anxiety are sleeping disturbances in patients. For this purpose, the candidate will work on real data from and in close collaboration with the local hospital.
Besides doing research, the candidate is expected to participate in educational activities (participate in seminars, be involved in teaching,…).
Profile
Offer
We offer full-time employment for a PhD researcher. Starting date is around 1 September 2020. The appointment will be for 1 year initially and will be extended after positive evaluation. Funding is available for 4 years, the typical duration of a PhD project. This position offers flexibility, the opportunity to work in a multicultural and multidisciplinary environment and to learn continuously. The candidate will be working at the Campus Kulak in Kortrijk. (S)he will enter the doctoral program of the KU Leuven Doctoral School of Biomedical Sciences.
In addition to a competitive scholarship, KU Leuven also offers a number of additional benefits, such as hospitalisation insurance, eco-vouchers, reimbursement of commuting expenses by public transport, inexpensive meals, a KU Leuven bicycle,…
Applications must be received no later than May 31, 2020.