Research Interests
My research interests are in the areas of Machine Learning and Artificial Intelligence in general. In particular, I am interested in all kinds of temporal settings, including learning from data streams and all that it implies (continual learning from dynamic and evolving data, transfer learning across tasks, and adapting to domain shift/concept drift — or robustness against it, in sequential data and time series) and also reinforcement learning and multi-agent systems.
I work often with methods of probabilistic machine learning (including Bayesian and Monte Carlo methods), and I focus on aspects such as uncertainty analysis, explainability, trustworthiness and reliability.
The above is relevant to many tasks such as predictive maintenance, diagnostics and forecasting, and industrial control systems. Indeed, I enjoy tackling real-world problems, and have worked and published in domains involving sensor networks and sensory data, complex transport and energy systems, medicine, as well as biology/ecology and the natural sciences.
I also still hold an long-term interest in multi-label classification, multi-target and structured-output prediction, particularly where they have merged into the modern context of deep learning, generative modelling, and sequential decision making.