If we want artificial agents to successfully support humans in everyday life, natural and smooth communication is the key. In this hands-on seminar, we will look at empirical evidence of how humans establish reference to objects in situated communication and what that means for human-machine interaction. We will examine current computational approaches for dealing with reference, and get a deeper understanding of them by exploring them practically in some exercises.
Based on the initial input on the topic, you will soon proceed to develop your own project on reference and grounding. This can be either an empirical study, or a system implementation.
No computer science or programming background is required for this course.