Multitask computation through dynamics in recurrent spiking neural networks

被引：0
作者：
Mechislav M. Pugavko
Oleg V. Maslennikov
Vladimir I. Nekorkin
机构：
[1] Institute of Applied Physics of the Russian Academy of Sciences,
来源：
Scientific Reports | / 13卷
关键词：
D O I：
暂无
中图分类号：
学科分类号：
摘要：
In this work, inspired by cognitive neuroscience experiments, we propose recurrent spiking neural networks trained to perform multiple target tasks. These models are designed by considering neurocognitive activity as computational processes through dynamics. Trained by input–output examples, these spiking neural networks are reverse engineered to find the dynamic mechanisms that are fundamental to their performance. We show that considering multitasking and spiking within one system provides insightful ideas on the principles of neural computation.
引用
收藏
相关论文
共 114 条
[1]  
Sporns O(2005)The human connectome: A structural description of the human brain PLoS Comput. Biol. 1 249-275
[2]  
Tononi G(2020)Computation through neural population dynamics Annu. Rev. Neurosci. 43 39-71
[3]  
Kötter R(2021)Cognitome: In search of fundamental neuroscience theory of consciousness Zhurnal Vysshei Nervnoi Deyatelnosti Imeni I.P Pavlova 71 1048-1070
[4]  
Vyas S(2020)Generative models of brain dynamics Front. Artif. Intell. 5 1089-1109
[5]  
Golub MD(2020)Artificial neural networks for neuroscientists: A primer Neuron 107 127-149
[6]  
Sussillo D(2022)Nonlinear dynamics and machine learning of recurrent spiking neural networks Physics-Uspekhi 192 2531-2560
[7]  
Shenoy KV(2009)Reservoir computing approaches to recurrent neural network training Comput. Sci. Rev. 3 350-794
[8]  
Anokhin K(2002)Real-time computing without stable states: A new framework for neural computation based on perturbations Neural Comput. 14 783-317
[9]  
Ramezanian-Panahi M(2016)Building functional networks of spiking model neurons Nat. Neurosci. 19 305-258
[10]  
Yang GR(2022)The role of population structure in computations through neural dynamics Nat. Neurosci. 25 94-6
12345678910