The prototype of a web- and cloud-based interface for labour monitoring and decision support, building on unsupervised dimensionality reduction, has been developed. It allows to monitor labour in real-time, as simplified trajectories, while comparing to a database of previous labour trajectories with known outcomes. This continuous comparison allows to dynamically update estimates of deviation from the expected complication-free trajectory as well as of chances of intervention or adverse outcomes. However, usability and interpretability from the point of view of the healthcare provider are still not optimised.

The main objective of this project is to improve these aspects of the interface, by processing the available information and providing it in simplified and useful manners, e.g. by means of scores, interactive plots/tables, alert codes, etc. A secondary objective would be exploring the potential of Gaussian Processes to improve the trajectory modelling component.

Preferred skills: comfortable with Matlab and python; notions of machine learning.