About the project
Time series machine learning is a rapidly evolving field of artificial intelligence research. This project involves researching new algorithms for time series classification, helping develop the open-source aeon python toolkit, and collaborating with international partners to apply these algorithms to applications in health technology and human activity recognition.
Time series classification (TSC) involves learning how to label really valued, ordered data. The Hierarchical Vote Collective of Transformation-based Ensembles (HIVE-COTE) framework is currently state of the art for TSC, but there is great scope for further improvements. The core principle of HIVE-COTE is that different discriminatory features in time series will be appropriate for different domains. For example, to learn how to identify an individual from their motion trace over time, we may want features derived from short segments that represent a characteristic motion, and we may not care when they occur.
We will extend the HIVE-COTE framework to achieve three core goals; improve the predictive performance; enhance efficiency and scalability; and facilitate greater usability and explainability. The successful candidate will join a vibrant research group consisting of two post docs and three PhD students. You will collaborate with the group to help perform case studies in human activity recognition and digital health. We all work with, contribute to, and help maintain a common open-source toolkit, aeon, which has become the research focus of groups around the world. There will be opportunities for collaborative research trips with world leading researchers in USA, Australia, France, Spain, Germany and Ireland and chances to attend top conferences in the field of machine learning and data mining. As part of the broader Vision, Learning and Control group you can engage in broader social and research activities.
Applications will be considered at any time, there is no fixed deadline.
