Round Corner
Department of Computer and Information Science


Measuring Prediction Model "Staleness" (AI Lab Pitch)

Description of Problem

Those making supervised Machine Learning models which make predictions on time series data make certain assumptions about the input data and about the stationarity of the systems they model. Such models can become invalid if these assumptions are breached. Knowing when a model becomes “stale” could be used to trigger retraining, or at the very least to flag that the model is operating in a degraded mode.

But what precisely does it mean for a model to be “appropriate” versus “stale”? How can this be measured, detected, and described in a useful way to those overseeing a model performance? We are interested in developing simple metrics for model appropriateness which provide concise comprehensible information through monitoring systems.

The project will use real data, and simulation data. Through artificial manipulation of the data to render the model inappropriate, the candidate will develop techniques for detecting the transition from appropriate to stale.

We have access to data from all our industries, including data from shipping (AIS), remote sensing, safety and incident monitoring systems and machinery condition monitoring. Which data sets will be used for this project will be determined based on both the candidate’s wishes and current project interests.

About Analytic Innovation Centre (AIC) at DNV GL

The Analytic Innovation Centre (AIC) at DNV GL is a team of data scientists and software experts with a variety of backgrounds from industry and academia working across DNV GL’s broad business areas: maritime, oil & gas, energy and business assurance. AIC works on machine learning, big data and algorithm assurance problems, and through our work seek to introduce common sense, risk-aware application of ML/AI technologies to a host of real-life problems.

DNV GL is a global quality assurance and risk management company. Driven by our purpose of safeguarding life, property and the environment, we enable our customers to advance the safety and sustainability of their business. We provide classification, technical assurance, software and independent expert advisory services to the maritime, oil & gas, power and renewables industries. We also provide certification, supply chain and data management services to customers across a wide range of industries. With origins stretching back to 1864 and operations in more than 100 countries, our experts are dedicated to helping customers make the world safer, smarter and greener.

Other Information

  • Time frame for thesis: 6 months
  • Contact: Justin Fackrell, Data Scientist, DNV GL, Høvik +47 46787123,


In this project, the joint interests of sponsor, advisor and student(s) come together. Typically, the project will be based on the problem described by the sponsor AND previous research by Prof. Ole Jakob Mengshoel AND interests of students. If only one or two of these are present, there is no basis for a project.

Please send email(s) to sponsor and/or advisor if you're interested in this project.

Ole Jakob Mengshoel is (more or less) following Keith Downing's selection process for master students - read this:


Ole Jakob Mengshoel Ole Jakob Mengshoel
313 IT-bygget
735 93440 
NTNU logo