Massimiliano Ruocco

This PhD project at the Department of Computer Science, Norwegian University of Science and Technology (NTNU) focuses on machine learning for temporal and spatiotemporal data , including time series and multivariate time series observed on graph nodes and edges. The research is centered on forecasting and on detecting disruptions in data drawn from complex networked systems. The position is embedded in STRIDE , an interdisciplinary NTNU project on the resilience of transport systems. Transport networks are the main testbed, and the work will study how learning methods can remain reliable when networks are pushed into rare or previously unseen operating regimes. The project aims to develop deep learning approaches that can both forecast under distribution shift and detect or locate disruptions from observed data. Possible research directions include time-series foundation models, pretraining on synthetic data from simulated dynamical systems, spatiotemporal graph neural networks that model how disruptions propagate through a network, and anomaly detection/localization in time series on graphs. The exact focus will be shaped together with the successful candidate, with an expectation of publication in leading machine learning conferences and journals. The PhD candidate will work in the Data and Artificial Intelligence (DART) group and collaborate closely with other researchers in the STRIDE team. The wider project connects transport modelling, artificial intelligence, operations research, civil engineering, and science and technology studies, creating an interdisciplinary environment around societal resilience and transport-system robustness. The main supervisor is Associate Professor Massimiliano Ruocco , with co-supervisors from the STRIDE team. The role includes research, publication, open release of code and benchmarks, participation in international activities, and one year of teaching or other career-promoting work at NTNU.