The National Research Council of Canada’s (NRC) Data Science for Complex Systems team works with industry and government to analyze data from complex engineering systems found in vehicles, buildings, utility grids, and advanced manufacturing settings. Using data about time and space gathered by sensors, we can model system performance, track system health, identify abnormal conditions, and predict failures.
We also conduct multidisciplinary research independently and in collaboration with academia to advance knowledge in machine learning and artificial intelligence toward long-term intelligence for computers. This includes the ability of artificial intelligence to explain its own machine learning and decision-making processes and bases, a field known as explainability that is growing exponentially as artificial intelligence becomes part of our daily lives.
What we offer
Based in the NRC's Digital Technologies Research Centre, the team's core competencies and techniques include:
- artificial intelligence
- case and rule-based reasoning
- classification and clustering
- collaborative filtering
- data mining and analysis
- deep learning
- inference engines
- information extraction
- machine learning
- natural language understanding
- numerical modelling
- operations research
- recommender systems
- rule-based systems
- statistical analysis
We apply these competencies to:
- materials design and development
- predictive maintenance, availability, and life cycle of engineering equipment and systems
- supply chain and transportation optimization
Our research also contributes to:
- deep space exploration systems
- health diagnostics
- workforce and learning management systems
Why work with us
Our experts have over 20 years of experience analyzing high-volume data from complex engineering systems and can provide insights that help industrial and government collaborators reduce equipment life cycle cost and increase equipment availability.
In terms of foundational research, we are interested in collaborating with groups who share the following research interests:
- Developing new algorithms for long term learning: dealing with changing concepts and transferring learning results from one task to another
- Harnessing human knowledge: exploiting human knowledge in formal models, updating or learning new models of similar form, explaining black box models, and speeding up the design process
Interested in harnessing our data science insights or conducting collaborative research with us? Contact our experts today!
- Long short-term memory over recursive structures.
- C4.5, Class Imbalance, and Cost Sensitivity: Why Under-Sampling beats OverSampling.
- Belacel, N. Multicriteria Assignment Method PROAFTN: Methodology and Medical Application. European Journal of Operational Research, (125), pp .175-83, 2000.
- Belanche L., Valdés J.J., Comas J., Roda I., Poch M. Prediction of the bulking phenomenon in wastewater treatment plants. Artificial Intelligence in Engineering, 14, (4), pp. 307-317, 2000.
- Guo, H., & Viktor, H.L. Learning from imbalanced data sets with boosting and data generation: the DataBoost-IM approach. SIGKDD Explorations, 6, pp. 30-39, 2004.