Enabling Advanced Autonomy through Human Interaction

The Universities of Edinburgh, West of England and Oxford are delighted to announce their new project Enabling Advanced Autonomy through Human-Collaboration, funded as part of Towards Turing 2.0 (second tranche). The project runs from 1 October 2021 until 31 September 2022.

Despite impressive advancements in AI, we are still a long way from systems that can perform complex activities independently in real-world domains – outside of controlled environments we are not able to build systems that exhibit the flexibility, adaptability and awareness required to perform those tasks which we would prefer AI to take on rather than humans.

In this subfield we will lay the foundations for new procedures, techniques and design tools  for AI systems that replicate a natural interaction. This will enable embodied AI agents such as social robots to learn new skills from human instructors through a continuous cycle of instruction, explanation, and feedback.

The project will focus on the following:

(1) semantics in conversational dialogue – piloted in the domain of teaching surgical assistance robots new skills; 

(2) to “learn” to construct models from real-world data – piloted on the problem of generating narratives from instructional videos

(3) learning the vocabulary and “grammar” humans apply when solving a task to enable AI systems to create human-understandable explanations of their own behaviours

The Responsible Technology Institute will be contributing to this project to provide expertise on human factors and ethical issues, together with the University of the West of England.

Find out more:https://blogs.ed.ac.uk/advanced-autonomy