Data SCIENCE & ARTIFICIAL INTELLIGENCE
"We combine the strength of both domain knowledge and data-driven approaches to develop strong, reliable methods. We extend our modelling with inverse models, trend predictions and highly non-linear regressions.' | Bart Mak
DATA-DRIVEN OPTIMISATION
We will enable fast, data-driven, multi-objective and operation-aware design exploration and optimisation in the early stages of ship development, overcoming current limitations in geometry definition, computational cost, and data availability.
DEEP LEARNING DATA ANALYSIS
We aim to advance deep learning capabilities from early-stage prototypes toward fully operational and integrated solutions that can reliably support complex maritime applications.
MLOPs
We want to develop a scalable, efficient, and collaborative machine learning ecosystem that ensures reliability, traceability, and effective integration of AI into engineering workflows. We will develop versioning of data-science models, their training data and hyperparameter settings, supported by tests with relevant benchmark models and
datasets.
REAL-TIME ENVIROMENTS
We will develop an advanced and integrated framework for real-time data-driven workflows to support development and analysis of on-board advisory and autonomous systems. The primary context is virtual environments, with realistic representations of data connections in operational contexts. This includes enhancing simulated traffic and (virtual) sensor data utilisation, enabling real-time decision support and supporting the next generation of AI-driven maritime applications.