Power-efficient neural network with artificial dendrites
Open Access
- 29 June 2020
- journal article
- research article
- Published by Springer Science and Business Media LLC in Nature Nanotechnology
- Vol. 15 (9), 776-782
- https://doi.org/10.1038/s41565-020-0722-5
Abstract
In the nervous system, dendrites, branches of neurons that transmit signals between synapses and soma, play a critical role in processing functions, such as nonlinear integration of postsynaptic signals. The lack of these critical functions in artificial neural networks compromises their performance, for example in terms of flexibility, energy efficiency and the ability to handle complex tasks. Here, by developing artificial dendrites, we experimentally demonstrate a complete neural network fully integrated with synapses, dendrites and soma, implemented using scalable memristor devices. We perform a digit recognition task and simulate a multilayer network using experimentally derived device characteristics. The power consumption is more than three orders of magnitude lower than that of a central processing unit and 70 times lower than that of a typical application-specific integrated circuit chip. This network, equipped with functional dendrites, shows the potential of substantial overall performance improvement, for example by extracting critical information from a noisy background with significantly reduced power consumption and enhanced accuracy.Keywords
Funding Information
- National Natural Science Foundation of China (61674089, 91964104, 61851404)
- Beijing Municipal Science and Technology Commission (Z191100007519008)
This publication has 53 references indexed in Scilit:
- A scalable neuristor built with Mott memristorsNature Materials, 2012
- Nonlinear dendritic processing determines angular tuning of barrel cortex neurons in vivoNature, 2012
- Locally Synchronized Synaptic InputsScience, 2012
- Short-Term Memory to Long-Term Memory Transition in a Nanoscale MemristorACS Nano, 2011
- Short-term plasticity and long-term potentiation mimicked in single inorganic synapsesNature Materials, 2011
- Dendritic Discrimination of Temporal Input Sequences in Cortical NeuronsScience, 2010
- The decade of the dendritic NMDA spikeJournal of Neuroscience Research, 2010
- Exponential ionic drift: fast switching and low volatility of thin-film memristorsApplied Physics A, 2008
- The distribution and targeting of neuronal voltage-gated ion channelsNature Reviews Neuroscience, 2006
- Cell Type-Specific Structural Plasticity of Axonal Branches and Boutons in the Adult NeocortexNeuron, 2006