WASP-funded postdoc position in Computer Science and Artificial Intelligence at Örebro University.
Background
The post-doc will be affiliated with the research Centre for Applied Autonomous Sensor Systems (AASS, http://www.oru.se/aass), which carries out multi-disciplinary research at the intersection of robotics, machine learning, artificial intelligence, computer vision, and computer science.
The project looks at how investigates how robots and autonomous systems can learn adaptively to successfully operate in changing environments. Real-world environments are constantly changing, and for a robot or autonomous system to operate successfully without human intervention or catastrophic failure, it must have both the capacity to recognise change and the flexibility to adapt its beliefs and behaviours. The goal of this project is to develop robust systems by which a robot can evaluate both its own models of the world, and the current environmental conditions, to provide confident predictive models.
The project investigates interpretable machine learning for vision-based applications. While machine learning is a powerful tool for the solution of complex problems, there is a growing focus on developing techniques that are interpretable and explainable, rather than relying on black-box models. Furthermore, there is widespread interest in learning algorithms that can discover or validate scientific hypotheses. In this project, the goal is to develop novel interpretability-based machine learning techniques for computer vision that not only predict decision outcomes from input data, but also can assist in relating the algorithmic output to the underlying properties of the systems under investigation.