OPTIMAL (Optimization for and with Machine Learning)
Machine learning has often made headlines in recent years, with spectacular applications such as image recognition, self-driving cars, and programs that beat the best human players at games like chess and Go. A key component of this technology is mathematical optimisation, that is used, for example, to train the underlying neural networks.
Our objective in this project is to develop innovative optimisation models, analysis techniques, and solution methods for a mutual reinforcement of optimisation and machine learning. We bring together a team of experts in mathematical optimisation, and a larger consortium that includes data scientists, who will study specific, interrelated, applications in machine learning. The goal is to provide new analysis and tools for optimisation problems and algorithms arising in machine learning, but also the other way around: to use insights and techniques from machine learning to improve optimisation models and methods. This explains the project title Optimisation for and with machine learning. The results from this project have a lot of potential to enhance both machine learning and optimisation. We will test our insights on a variety of applications where the consortium members are already involved, including routing of shared, self-driving cars, classification problems in the medical sciences, and decision problems related to the UN World Food Programme.
Please find the facts and figures of the OPTIMAL Optimisation for and with Machine Learning research project:
To the NWO press release (February 25, 2020).