The confluence of advances in diverse areas including machine learning, large scale computing and reliable commoditised hardware have brought autonomous robots to the point where they are poised to be genuinely a part of our daily lives. Application areas where these autonomous robots must interact with human co-workers also bring with them stringent requirements regarding safety, explainability and trustworthiness. These needs seem to be at odds with the ways in which recent AI successes have been achieved, e.g., with end-to-end learning.

In this talk, I will describe key elements of these recent advances, using as examples robot systems developed in our lab. We will look at systems that are capable of intention-aware navigation in human environments, collaborative work with human co-workers and the use of dialogue to faciliate task specification by the layperson. A common theme underpinning these systems is the use of machine learning techniques, ranging from Bayesian inference to deep generative models.

I will end by describing how researchers are beginning to address the gaps between powerful learning methods and understandably safe systems, through the use of structured representations.