Stanford And Oxford Researchers Propose An Approach To Relate Transformers To Models And Neural Representations Of The Hippocampal Formation

In recent years, a significant part of neuroscience research has focused on relating deep learning architectures to the human brain, and many deep learning (DL) techniques have recently been shown to replicate neural firing patterns observed in the brain. For example, representations of convolutional neural networks have been shown to predict neurons in the visual cortex and inferior temporal cortex, while recurrent neural networks have been shown to recapitulate grid cells in the medial entorhinal cortex. The ability to use …

artificial oxford researchers stanford transformers