This project will involve up to 15 robotics and vision, natural language processing, knowledge representation, multiagent systems, cognitive science, and human factors researchers working together to demonstrate the potential of novel, integrative AI methods to advance autonomy.

Focusing specifically on new forms of human-AI collaboration through naturalistic interaction as a new design methodology, we will work on enabling embodied AI agents to learn new skills from human instructors at run-time through a continuous cycle of instruction, explanation, and feedback.

Alongside this core technical research, we will also explore the human and ethical dimensions of establishing advanced levels of autonomy through new forms of human-AI collaboration.

Researchers: Michael Rovatsos (Principal Investigator), Oisin Mac Aodha, Ram Ramamoorthy, Laura Sevilla, Alan Winfield, Hakan Bilen, Bob Fisher, Marina Jirotka, Alex Lascarides, Frank Keller, Anouk van Maris, Frank Mollica, Sid Narayanaswamy, Pericle Salvini, Antonio Vergari

The project builds on many ideas developed as part of the emerging Edinburgh Laboratory for Integrated AI led by Mirella Lapata as part of her Turing AI World-Leading Researcher Fellowship, which involves a host of additional AI researchers and industry partners, and pursues a complementary focus, which is the integration of different machine learning and symbolic reasoning modules to enable human-like reasoning in AI systems.