Cognitive AI
Benefits from (or provides insight into) the human condition
Programme Information
Human-like computing aims to equip machines with human-like learning, reasoning and perceptual abilities. Such abilities can help computers to interpret the aims and intentions of humans and they can enable better collaboration and communication between humans and machines. Human-like computing may integrate knowledge from AI, cognitive, biological and other sciences. It can ultimately benefit many applications that require close interaction between computers and human users, e.g., AI-powered tutoring, personalised healthcare, among others.
AI for neuroscience is another relevant theme – one that uses a variety of techniques (including computational, imaging and other techniques) to study the human brain and cognitive abilities. These may include improved understanding of the functionality and development of the human brain over the lifespan, as well as conditions such as autism, dementia and mental disorders. Research in cognitive AI combines knowledge and techniques from e.g., cognitive, computational, biological and clinical sciences.
Featured
Deep Cognition Lab
The Deep Cognition Lab uses cognitive computational models to try to understand high-level cognition and investigate methods to constrain computational models of cognition by brain data. The current focus is on using ECoG and speech data to map concepts in the neuronal space.
Lab Lead: Ariel Goldstein
CAPAIBLE (Cognition And Predictability in AI Behaviour Laboratory for Evaluation)
CAPAIBLE studies the cognitive capabilities of AI systems using methods from cognitive science and measurement theory. Their work spans projects such as Animal AI, which explores embodied cognition through comparative approaches, and the Future of Skills Project, which aligns AI capabilities with human job requirements. The team map cognitive profiles of AI systems to better understand their behaviour.
Lab Leads: Lucy Cheke & José Hernández-Orallo
Adaptive Brain Lab
The Adaptive Brain Lab uses an interdisciplinary approach that combines behavioural paradigms, movement recording, multimodal brain imaging (MRI, EEG, MEG, TMS) and state-of-the-art computational methods to learn more about how humans perceive the structure of the world around them.
Lab Lead: Zoe Kourtzi
