Artificial neural networks(ANN) have been extensively researched due to their significant energy-saving benefits.Hardware implementations of ANN with dropout function would be able to avoid the overfitting problem. Th...Artificial neural networks(ANN) have been extensively researched due to their significant energy-saving benefits.Hardware implementations of ANN with dropout function would be able to avoid the overfitting problem. This letter reports a dropout neuronal unit(1R1T-DNU) based on one memristor–one electrolyte-gated transistor with an ultralow energy consumption of 25 p J/spike. A dropout neural network is constructed based on such a device and has been verified by MNIST dataset, demonstrating high recognition accuracies(> 90%) within a large range of dropout probabilities up to40%. The running time can be reduced by increasing dropout probability without a significant loss in accuracy. Our results indicate the great potential of introducing such 1R1T-DNUs in full-hardware neural networks to enhance energy efficiency and to solve the overfitting problem.展开更多
Neuromorphic devices that mimic the information processing function of biological synapses and neurons have attracted considerable attention due to their potential applications in brain-like perception and computing. ...Neuromorphic devices that mimic the information processing function of biological synapses and neurons have attracted considerable attention due to their potential applications in brain-like perception and computing. In this paper,neuromorphic transistors with W-doped In_(2)O_(3)nanofibers as the channel layers are fabricated and optoelectronic synergistic synaptic plasticity is also investigated. Such nanofiber transistors can be used to emulate some biological synaptic functions, including excitatory postsynaptic current(EPSC), long-term potentiation(LTP), and depression(LTD). Moreover, the synaptic plasticity of the nanofiber transistor can be synergistically modulated by light pulse and electrical pulse.At last, pulsed light learning and pulsed electrical forgetting behaviors were emulated in 5×5 nanofiber device array.Our results provide new insights into the development of nanofiber optoelectronic neuromorphic devices with synergistic synaptic plasticity.展开更多
基金Project supported by the National Key Research and Development Program of China (Grant Nos. 2021YFA1202600 and 2023YFE0208600)in part by the National Natural Science Foundation of China (Grant Nos. 62174082, 92364106, 61921005, 92364204, and 62074075)。
文摘Artificial neural networks(ANN) have been extensively researched due to their significant energy-saving benefits.Hardware implementations of ANN with dropout function would be able to avoid the overfitting problem. This letter reports a dropout neuronal unit(1R1T-DNU) based on one memristor–one electrolyte-gated transistor with an ultralow energy consumption of 25 p J/spike. A dropout neural network is constructed based on such a device and has been verified by MNIST dataset, demonstrating high recognition accuracies(> 90%) within a large range of dropout probabilities up to40%. The running time can be reduced by increasing dropout probability without a significant loss in accuracy. Our results indicate the great potential of introducing such 1R1T-DNUs in full-hardware neural networks to enhance energy efficiency and to solve the overfitting problem.
基金Project supported by the National Key Research and Development Program of China (Grant Nos. 2021YFA1200051 and 2019YFB2205400)the National Natural Science Foundation of China (Grant Nos. 62174082 and 62074075)。
文摘Neuromorphic devices that mimic the information processing function of biological synapses and neurons have attracted considerable attention due to their potential applications in brain-like perception and computing. In this paper,neuromorphic transistors with W-doped In_(2)O_(3)nanofibers as the channel layers are fabricated and optoelectronic synergistic synaptic plasticity is also investigated. Such nanofiber transistors can be used to emulate some biological synaptic functions, including excitatory postsynaptic current(EPSC), long-term potentiation(LTP), and depression(LTD). Moreover, the synaptic plasticity of the nanofiber transistor can be synergistically modulated by light pulse and electrical pulse.At last, pulsed light learning and pulsed electrical forgetting behaviors were emulated in 5×5 nanofiber device array.Our results provide new insights into the development of nanofiber optoelectronic neuromorphic devices with synergistic synaptic plasticity.