PhD Candidate in Machine Learning for Temporal and Spatiotemporal Data

This is NTNU NTNU is a broad-based university with a technical-scientific profile and a focus in professional education. The university is located in three cities with headquarters in Trondheim. At NTNU, 8,6000 employees and 44,000 students work to create knowledge for a better world. You will find more information about working at NTNU and the application process here. Video: https://www.youtube.com/watch?v=_KHQjc4ndas&t=41s About the position The Department of Computer Science (IDI) at NTNU in Trondheim has a vacancy for a PhD Candidate in machine learning (four-year position). The topic of the PhD is machine learning for temporal and spatiotemporal data, that is, time series, including multivariate time series observed on the nodes and edges of a graph, with a focus on forecasting and on detecting disruptions. The position is part of STRIDE, an interdisciplinary NTNU project on the resilience of transport systems (see About the project), and transport networks are the main testbed. Forecasting models are usually trained on data from normal operating conditions. When a disruption, such as a closed link, a capacity limit or a sudden change in demand, moves a network into a regime that is rare or absent in the training data, their performance degrades, and such events are often too rare to learn from directly. The PhD will develop deep learning methods that forecast reliably under this kind of distribution shift, and that can detect and locate disruptions from the observed data. Possible research directions include: time-series foundation models, including pretraining on synthetic data from simulated dynamical systems, to handle events that are rare in real data spatiotemporal graph neural networks that model how disruptions propagate through a network anomaly detection and localization of disruptions in time series on graphs The exact research focus will be defined together with the candidate. We expect the results to be published at leading machine learning conferences and journals. The main supervisor will be Associate Professor Massimiliano Ruocco, with co-supervisors from the STRIDE team. Are you motivated to take a step towards a doctorate and open up exciting career opportunities? As a PhD Candidate with us, you will work to achieve your doctorate, and at the same time gain valuable experience that qualifies you for a further career in higher education and research, in and outside academia. Your immediate leader will be the Head of DART (Data and...

Summary and match are AI-generated; verify details on the original posting.

Listed via EURES. All sources