Testing a deep-neural network model of cognitive map development

测试认知图开发的深度神经网络模型

• 批准号: 2725859
• 金额: --
• 依托单位: University College London
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2022
• 资助国家: 英国
• 起止时间: 2022 至 无数据
• 项目状态: 未结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-2725859
• 关键词: Testing; deep; neural; network; model

The brain contains a 'cognitive map' of space, based on the firing of neurons in thehippocampus tuned to an animal's position and orientation, which can remember goallocations and navigate towards them. Artificial navigating agents based on deeplearningnetworks exhibit similar spatially-tuned neurons, however, it is not known howthese networks develop as the agents are trained, and whether this reflectsbiologically realistic learning and network development. This project combines deepneuralnetwork modelling with in vivo electrophysiological recording from juvenilerodents, aiming to test whether deep networks can model the biological developmentof a neural network for navigation.

大脑包含一个关于空间的“认知地图”，它基于海马体中的神经元的放电，这些神经元被调整到动物的位置和方向，可以记住目标分配并向它们导航。基于深度学习网络的人工导航代理表现出类似的空间调谐神经元，然而，目前还不知道这些网络如何随着代理的训练而发展，以及这是否反映了生物学上的现实学习和网络发展。该项目将深度神经网络建模与青少年的体内电生理记录相结合，旨在测试深度网络是否可以模拟神经网络的生物发育以进行导航。