The technical side of sense-making in social and other robots
Recording from AI Lund lunch seminar 31 August 2022
Topic: The technical side of sense-making in social and other robots
Speaker: Elin Anna Topp, Senior lecturer, Robotics and Semantic Systems, Computer Science, Lund University
When: 31 August at 12.00-13.15
Where: Online
Spoken language: English
Abstract
This presentation aspires to be somewhat of a counterpart or response to the previous AI Lund Lunch seminar held by Valentina Fantasia, in which a case for a paradigm shift in both human-human and human-robot interaction (HRI) research was made, to be able to provide genuinely engaging social interaction qualities in (social) robots.
Just as in this earlier presentation, I will show, why this paradigm shift was and is necessary not only in social robotics, but how also service and even industrial robots - and their users - can benefit from mutual, genuine understanding, i.e., from really making sense of each other’s actions and utterances.
To this end, I will discuss actual technical approaches to attempted sense-making - both failing and succeeding such - that I have encountered during various user studies and system evaluations. To some extent, I take the robotics perspective rather than the human perspective to explain the technical aspects of HRI that are needed to achieve smooth interaction between robots and humans, and I will make (repeat) an argument for why looking at failures and even more so at what we could call silent failures is more urgent than looking at providing ever better individual cognitive functionalities in robots.