AI model unveils how memories blend reality and imagination

Recent UCL study explores the role of generative AI in explaining how memories enable learning, imagination, and planning. The study uses a generative neural network to simulate how the brain’s neural networks learn from and remember events. The hippocampus and neocortex work together during memory, imagination, and planning. The AI networks suggest that replaying memories during rest helps the brain pick up on patterns from past experiences for making predictions. The neocortical network learns efficient conceptual representations of scenes, enabling the recreation of old scenes and generation of new ones. This explains how the brain slowly acquires conceptual knowledge and reconstructs events in the mind, as well as why existing memories often contain distortions.

Source link