Study Finds Striking Similarities Between AI and Human Memory, Opening Path to More Efficient Brain-Like Systems
-
Researchers found striking similarity between AI memory processing and human hippocampal functions, bridging neuroscience and AI.
-
The study reveals Transformer model in AI uses a gating process like the NMDA receptor in the brain, crucial for memory consolidation.
-
This finding allows developing more efficient, brain-like AI systems and better understanding of human memory.
-
Tweaking Transformer model to reflect NMDA receptor's role enhanced its memory, like controlling magnesium levels affects brain memory strength.
-
Research convergence of human cognition and AI advances AGI development and provides insights into brain through AI models.