Deep learning and the Global Workspace Theory

Recent advances in deep learning have allowed artificial intelligence (AI) to reach near human-level performance in many sensory, perceptual, linguistic, and cognitive tasks. There is a growing need, however, for novel, brain-inspired cognitive architectures. The Global Workspace Theory (GWT) refers to a large-scale system integrating and distributing information among networks of specialized modules to create higher-level forms of cognition and awareness. We argue that the time is ripe to consider explicit implementations of this theory using deep-learning techniques. We propose a roadmap based on unsupervised neural translation between multiple latent spaces (neural networks trained for distinct tasks, on distinct sensory inputs and/or modalities) to create a unique, amodal Global Latent Workspace (GLW). Potential functional advantages of GLW are reviewed, along with neuroscientific implications.

Automated summary

Recent advances in deep learning have enabled artificial intelligence to perform nearly as well as humans in sensory, perceptual, linguistic, and cognitive tasks, sparking a need for brain-inspired cognitive architectures. The Global Workspace Theory proposes a large-scale system for integrating and distributing information to specialized modules, creating higher-level forms of cognition and awareness. Implementing this theory using deep-learning techniques through unsupervised neural translation can potentially offer functional advantages and have significant implications in neuroscience.