Research focus is on machine learning, computational learning theory, natural language processing and generation, story comprehension, symbolic knowledge representation, commonsense reasoning, formal argumentation, cognitive computing, explainable AI.
Machine Learning is huge right now, especially in relation to the processing of visual and textual inputs. A recently emerging area within Machine Learning is Explainable AI, which seeks to build machines that can explain their decisions to ordinary humans. Universities, businesses, and governmental bodies across the world are working towards this direction.
This MRG seeks to complement these efforts in Explainable AI through the development of socially-competent agents that naturally interact and explain themselves to their human collaborators.
The MRG focuses on the design and implementation of protocols that support:
A machine’s ability to acquire knowledge about social interactions and norms, through the observation of, and interaction with, humans in social contexts.
The machine-aided completion by humans of everyday tasks, and the machine-aided formation of competent and effective human teams for specified goals and tasks.
A machine’s ability to engage in a dialectical interaction with other machines or humans, and explain its reasoning behind a certain inference or action, while also being able to improve by accepting feedback from humans on the quality of its chosen decisions.
The MRG will undertake research at the cutting edge of modern AI research and innovation, combining machine learning, reasoning and argumentation frameworks, multi-agent models, and human-machine interaction. The recent proliferation of the use of such AI techniques across areas and domains ensures that the MRG will be uniquely positioned to have applications in a wide range of fields of human activity that would stand to benefit from the use of physical, virtual, or software assistants. The MRG will interact with other groups within RISE with expertise in: human-centred design, so that appropriate human factors are taken into account in the development of the protocols; the undertaking of experimental work with human participants, and especially on how to establish the cognitive-compatibility of the developed protocols with humans in the context of mixed machine-human teams; the processing of visual sensory inputs, and the rendering of virtual worlds, through which machines will be able to interact with humans without requiring physical proximity; the communication with other physical machines or smart devices, which will allow a machine to enhance its awareness of, and accessibility to, the environment that it populates and within which team members reside.
Dr. Loizos Michael