Theory of mind at the interface of neuroscience and AI
The Visual AI Laboratory has secured, in partnership with the University of Cambridge, a Leverhulme Trust Research Project Grant entitled "Theory of mind at the interface of neuroscience and AI".
The duration of the project is 30 months, starting September 2019.
The Cambridge co-investigator is Professor Barbara Sahakian, a leading figure in the field of neuroscience with more than 60,000 scholar citations.
Emerging applications of artificial intelligence are highlighting the limitations of established approaches in situations involving humans. The integration of neuroscience and machine learning has the potential to enable significant advances in both fields. Theory of Mind capabilities, i.e., the ability to 'read' other sentient beings' mental states, are crucial for the development of a next generation, "human-centric" artificial intelligence aimed to understand the behaviour of complex agents. In a mutually beneficial process, computational models developed within artificial intelligence could provide new insights about how these mechanisms work in the human brain.
A committed philanthropist from the beginning, on his death in 1925 Lord Leverhulme left a share of his holdings in his company to provide for specific trades charities, and to offer ‘scholarships for… research and education’. The Leverhulme Trust was established to undertake these charitable aims. In 1930, Lever Brothers merged with Margarine Unie to form Unilever – one of the world’s major multinational companies – and the shares held by the Leverhulme Trust became shares in Unilever PLC.
The University of Cambridge is a collegiate research university in Cambridge, United Kingdom. Founded in 1209 and granted a royal charter by Henry III in 1231, Cambridge is the second-oldest university in the English-speaking world and the world's fourth-oldest surviving university.
Our fundamental hypothesis is that the integration of neuroscience and machine learning could enable significant advances in both fields, by allowing: (i) within neuroscience, the formulation and empirical validation of computational Theory of Mind models in humans leveraging current frontier efforts in ML and AI, forging towards a more detailed understanding of these functionalities in the human brain; (ii) within artificial intelligence, the development of machine Theory of Mind models informed by the latest neuroscientific studies and evidence. Our stance is that a simulation standpoint, which describes human ability to understand other people as the result of an internal ‘simulation’ of their mental processes, is the most suitable to bring together frontier machine learning and neuroscience in a radically new approach to the problem. Efficient simulations of human behaviour can be implemented using reconfigurable (deep) neural networks, able to change topology to adapt to different agent classes and scenes.
Success will lay the foundations for the creation, among others, of autonomous vehicles able to negotiate complex road situations involving humans. Next-generation robotic assistant surgeons can be envisaged with the ability to understand what the main surgeon is doing and foresee their future intentions, in order to best assist them. Empathic healthcare is becoming a priority for the NHS, especially when dealing with autism and similar conditions. The new theory of mind models we propose may improve the efficacy of psychological treatments, such as cognitive behavioural therapy or mindfulness. A new generation of ‘bots’ (e.g. for customer service or financial advice) able to interact more effectively and empathetically with humans would be possible. This work would also impact on the current debate on moral AI, helping machines make ethical, human-like decisions in critical situations. In the longer term robotic companions for disabled people can be imagined, based on the capabilities we aim to develop.