Making sense of a streaming world
Much of today's information is generated continuously. Traffic cameras track how busy the roads are, and drones overhead emit telemetry, much as servers on the ground do. Emerging technologies such as collaborative autonomous systems both produce and consume information, drawing on external streams to carry out their own tasks. These stream-generating resources are increasingly heterogeneous, geographically distributed, and unreliable, with no centralised ownership. Yet the streams they produce are critical to a better understanding of the world—an understanding that in turn strengthens democracy, encourages innovation, and creates value. No single person can keep track of what information exists, let alone how to access it. And because the landscape of streams is constantly shifting, even maintaining the streams needed to answer a single question is difficult. I develop artificial intelligence techniques to address these problems. My research centres on expressing spatio-temporal questions, finding the information needed to answer them, negotiating access across organisations, and robustly orchestrating the resulting processing—all while keeping auditability in mind.
Applied AI education
AI is a fast-moving field: data-driven approaches have produced major advances in areas such as computer vision, reinforcement learning, and natural language processing. Agentic large language models add powerful new tools, full of promise but also raising fresh challenges—including for education itself. I am involved in several initiatives to build on the strong AI education already offered at Linköping University. I helped establish AI Academy, where I also provide technical support; it sets practical AI challenges for talented students across all programmes and runs upskilling courses for industry partners and government organisations. I am also responsible for developing an arena for applied AI education, offering courses that combine data-driven AI, its regional and organisational context, and the hands-on use of AI tools.
