[Paper Review] On the link between conscious function and general intelligence in humans and machines

Arthur Juliani, Kai Arulkumaran+2 more|arXiv (Cornell University)|Mar 24, 2022

Functional Brain Connectivity Studies21 citations

TL;DR

The paper analyzes three functional theories of consciousness (Global Workspace Theory, Information Generation Theory, and Attention Schema Theory) to argue their link to domain-general intelligence and discusses how AI can integrate these concepts to pursue mental time travel as a path to stronger general intelligence.

ABSTRACT

In popular media, there is often a connection drawn between the advent of awareness in artificial agents and those same agents simultaneously achieving human or superhuman level intelligence. In this work, we explore the validity and potential application of this seemingly intuitive link between consciousness and intelligence. We do so by examining the cognitive abilities associated with three contemporary theories of conscious function: Global Workspace Theory (GWT), Information Generation Theory (IGT), and Attention Schema Theory (AST). We find that all three theories specifically relate conscious function to some aspect of domain-general intelligence in humans. With this insight, we turn to the field of Artificial Intelligence (AI) and find that, while still far from demonstrating general intelligence, many state-of-the-art deep learning methods have begun to incorporate key aspects of each of the three functional theories. Having identified this trend, we use the motivating example of mental time travel in humans to propose ways in which insights from each of the three theories may be combined into a single unified and implementable model. Given that it is made possible by cognitive abilities underlying each of the three functional theories, artificial agents capable of mental time travel would not only possess greater general intelligence than current approaches, but also be more consistent with our current understanding of the functional role of consciousness in humans, thus making it a promising near-term goal for AI research.

Motivation & Objective

• Assess how three contemporary theories of conscious function map onto domain-general intelligence in humans.
• Examine how current AI systems incorporate aspects of GWT, IGT, and AST to improve generalization.
• Propose a unified, implementable model inspired by these theories to enable artificial mental time travel.

Proposed method

• Review and compare Global Workspace Theory (GWT), Information Generation Theory (IGT), and Attention Schema Theory (AST) as accounts of conscious access.
• Discuss the relationship between conscious access and domain-general cognitive abilities.
• Survey how state-of-the-art deep learning approaches already integrate elements of each theory.
• Introduce the concept of mental time travel as a guiding example for integrating the theories into AI.

Experimental results

Research questions

• RQ1 How are GWT, IGT, and AST each linked to domain-general intelligence in humans?
• RQ2 In what ways have modern AI systems begun to implement aspects of GWT, IGT, and AST?
• RQ3 Can a unified, implementable model combining these theories enable artificial agents to perform mental time travel?
• RQ4 What are the cognitive mechanisms (attention, generative modeling, high-level meta-models) that could underpin more general AI?

Key findings

• All three theories tie conscious function to aspects of domain-general intelligence.
• Current AI methods already incorporate elements of GWT, IGT, and AST to achieve greater generalization.
• A unified model leveraging selective attention, a generative cognitive map, and an attention policy could enable mental time travel in agents.
• Mental time travel could yield AI with broader generalization than present approaches.