Representational Dynamics of Learning in Humans and Machines
Lead Research Organisation:
University of Cambridge
Department Name: Computer Science and Technology
Abstract
Trained deep neural networks are predictive of human behaviour and associated brain responses.
Through training, these networks learn to transform and represent information in order to facilitate task performance, in ways which can be similar to the brain.
Using the framework of visual perceptual learning, this project uses the training of neural networks to model and explain human learning. Using both behavioural and fMRI data, we first use these networks to build predictive models. Second, we interrogate these models for insight into the mechanisms of sensory learning, neuronal information representation, and decision making.
Through training, these networks learn to transform and represent information in order to facilitate task performance, in ways which can be similar to the brain.
Using the framework of visual perceptual learning, this project uses the training of neural networks to model and explain human learning. Using both behavioural and fMRI data, we first use these networks to build predictive models. Second, we interrogate these models for insight into the mechanisms of sensory learning, neuronal information representation, and decision making.
