Model Building

Team Lead: Ken Brown

Models are central components of statistics, data science, artificial intelligence and optimisation. The model describes the structure and relationships between different features of the data and the problem, and is used to describe observed patterns, predict future outcomes, or constrain recommended actions. Different choices of model type influence the accuracy of our predictions, and can enhance or limit our ability to explain or validate the decisions we make.

Building these models automatically from data is a major research challenge. In Insight, we are developing new techniques for efficient construction of large models from vast and complex data sets, for constructing models that make it easy to extract intuitive explanations, for maintaining the privacy of personal data while still allowing useful inference, and for generating decision and optimisation models automatically from observations.