The Maritime Research Institute the Netherlands (MARIN) performs research on ships through model tests, numerical simulations and through in-service measurements. Physics-based models are used to describe vessel behaviour and the measurements are used to confirm these models and gain additional insight in the vessel behaviour.
The Big data evolution has stimulated development in data collection, management, infrastructure and analysis technology, which may be of interest to MARIN's day-to-day analysis and new data driven analysis methods, including Machine Learning and Artificial Intelligence. These technologies will also prove crucial in the challenges ahead for the maritime shipping industry, such as the adoption of unmanned shipping.
MARIN is looking for a professional data science student to develop and apply Machine Learning and Artificial Intelligence technology to one of its existing data sets. The goal will be to obtain non-trivial information from data sets.
Your objective is to apply Machine Learning and Artificial Intelligence technology to perform data mining on a selected set of in-service measurements.
Together with your supervisor you will explore the data and gain understanding of the measurements. You will examine the possibility to include known physical relations between the parameters and provide a database of enhanced measurements. Another important item is the generalization of your findings. Are the observed relations are applicable to multiple applications or are they vessel dependent?
In some cases comparison between model tests, numerical simulations and in-service measurements may be possible. In that case, your work may be used to test physical scaling laws, or be used for development of numerical analysis techniques.
To conclude, we foresee the following tasks:
Gain insight in the available data set and gain understanding of the physical relevance of the measurements and their uncertainties.
If required, place the data in a format that works well with the available analysis tools.
Clean the data by removing trivial, non-physical and well-known relations.
Apply data mining to discover non-trivial internal relations in the data set.
Test the observed relations by comparison of multiple data sets, or through comparison between different scales or using numerical analysis.
The internship is for the duration of 6 to 9 months. The start date of the internship period can be determined in consultation with the supervisor.
For these tasks a data scientist student with affinity for engineering is foreseen. You should be willing to invest in understanding the relevant physics to obtain the highest added value possible. Subject matter experts from MARIN will be available to support your data interpretation.
Department and supervisor
During the internship you will be connected with the Trials & Monitoring (T&M) department of MARIN but cooperate with several employees throughout MARIN. The supervisor for this internship will be Remco Hageman.
Solliciteer / Apply
Internship: “Data science”