Advancing cognitive science: AI-inspired insights into long-term working memory and thinking

In the article "Thinking about thinking: AI offers theoretical insights into human memory," Terrence Sejnowski, Francis Crick Chair at the Salk Institute for Biological Studies, explores the need for a new conceptual framework to understand cognitive functions, particularly how the brain forms and maintains globally distributed states over extended periods. Drawing parallels between human cognition and artificial intelligence (AI), Sejnowski proposes that mechanisms like traveling waves and spike-timing-dependent plasticity (STDP) in the brain, inspired by AI’s generative transformers, could explain long-term working memory and cognitive processes such as thinking and planning.
If you're developing brain-inspired semiconductor processor chips, this article offers critical insights into the mechanisms of human cognition. It explores how traveling waves and spike-timing-dependent plasticity (STDP) in the brain, inspired by AI's generative transformers, support long-term working memory and complex cognitive processes like thinking and planning, providing a valuable framework for designing neuromorphic computing systems.

Sejnowski begins by highlighting the limitations of current cognitive science frameworks, which are better suited for rapid sensorimotor actions (lasting seconds) than for complex cognitive processes like reading, language comprehension, and plan...