Improved Conductance Linearity and Conductance Ratio of 1T2R Synapse Device for Neuromorphic Systems

We report on a 1-transisor/2-resistor (1T2R) synapse device with improved conductance linearity and conductance ratio under an identical pulse condition for hardware neural networks with high pattern-recognition accuracy. Utilizing an additional series-connected resistor, the conductance linearity of a synapse device was significantly improved owing to the reduced initial voltage drop on an resistive RAM (RRAM) device during depression conditions. Moreover, to maximize the conductance ratio of a synapse device, we utilized a steep subthreshold region of an MOSFET by a parallel connection of an RRAM and a transistor. A small change in voltage on the RRAM directly controlled the gate bias of the MOSFET, which causes a large change in the drain current. Compared with a conventional RRAM synapse device, the 1T2R synapse device shows an improved conductance linearity and conductance ratio ( $> \times 100$ ). Finally, we confirmed an excellent classification accuracy by using a neural network simulation based on a multilayer perceptron.

[1]  Hyunsang Hwang,et al.  Analog Synapse Device With 5-b MLC and Improved Data Retention for Neuromorphic System , 2016, IEEE Electron Device Letters.

[2]  Jacques-Olivier Klein,et al.  Spin-Transfer Torque Magnetic Memory as a Stochastic Memristive Synapse for Neuromorphic Systems , 2015, IEEE Transactions on Biomedical Circuits and Systems.

[3]  Yoon-Ha Jeong,et al.  Optimization of Conductance Change in Pr1–xCaxMnO3-Based Synaptic Devices for Neuromorphic Systems , 2015, IEEE Electron Device Letters.

[4]  Shimeng Yu,et al.  Scaling-up resistive synaptic arrays for neuro-inspired architecture: Challenges and prospect , 2015, 2015 IEEE International Electron Devices Meeting (IEDM).

[5]  Simone Balatti,et al.  A 2-transistor/1-resistor artificial synapse capable of communication and stochastic learning in neuromorphic systems , 2015, Front. Neurosci..

[6]  Pritish Narayanan,et al.  Experimental Demonstration and Tolerancing of a Large-Scale Neural Network (165 000 Synapses) Using Phase-Change Memory as the Synaptic Weight Element , 2014, IEEE Transactions on Electron Devices.

[7]  Dhireesha Kudithipudi,et al.  Comparison of Off-Chip Training Methods for Neuromemristive Systems , 2015, 2015 28th International Conference on VLSI Design.

[8]  G. W. Burr,et al.  Experimental demonstration and tolerancing of a large-scale neural network (165,000 synapses), using phase-change memory as the synaptic weight element , 2015, 2014 IEEE International Electron Devices Meeting.

[9]  Farnood Merrikh-Bayat,et al.  Nonlinear behavior of memristive devices during tuning process and its impact on STDP learning rule in memristive neural networks , 2014, Neural Computing and Applications.

[10]  Yukihiro Kaneko,et al.  Ferroelectric Artificial Synapses for Recognition of a Multishaded Image , 2014, IEEE Transactions on Electron Devices.

[11]  J. Kim,et al.  Neuromorphic speech systems using advanced ReRAM-based synapse , 2013, 2013 IEEE International Electron Devices Meeting.

[12]  O. Cueto,et al.  Physical aspects of low power synapses based on phase change memory devices , 2012 .

[13]  X. Bai,et al.  Categorization of resistive switching of metal-Pr0.7Ca0.3MnO3-metal devices , 2009 .