In this paper, we reviewed the recent trends on neuromorphic computing using emerging memory technologies. Two representative learning algorithms used to implement a hardware-based neural network are described as a bio-inspired learning algorithm and software-based learning algorithm, in particular back-propagation. The requirements of the synaptic device to apply each algorithm were analyzed. Then, we reviewed the research trends of synaptic devices to implement an artificial neural network.
A reconfigurable field-effect transistor (RFET) with memory functionality is proposed as a new high-density synaptic device capable of exclusive NOR (XNOR) operation for a binary neural network (BNN). A RFET with three gates is used in this letter, consisting of two program gates (PGs) electrically connected to each other and a control gate (CG) between the PGs. By changing the polarity of the PG bias or the polarity of the charge trapped in Si 3 N 4 layer on the PGs, the RFET operates as an nor a p-MOSFET having a CG as a switching gate. The XNOR operation is successfully demonstrated in a fabricated RFET. The on/off current ratio (the ratio of 1 to 0 in the result of the XNOR operation) is ~10 4 and the current is ~1 nA/μm. Using a RFET as a synaptic device for a BNN is efficient in terms of the area (~8F 2 /synapse) because XNOR operation can be performed using only one RFET.
Neuromorphic computing (NC) architecture inspired by biological nervous systems has been actively studied to overcome the limitations of conventional von Neumann architectures. In this work, we propose a reconfigurable NC block using a flash-type synapse array, emerging positive feedback (PF) neuron devices, and CMOS peripheral circuits, and integrate them on the same substrate to experimentally demonstrate the operations of the proposed NC block. Conductance modulation in the flash memory enables the NC block to be easily calibrated for output signals. In addition, the proposed NC block uses a reduced number of devices for analog-to-digital conversions due to the super-steep switching characteristics of the PF neuron device, substantially reducing the area overhead of NC block. Our NC block shows high energy efficiency (37.9 TOPS/W) with high accuracy for CIFAR-10 image classification (91.80%), outperforming prior works. This work shows the high engineering potential of integrating synapses and neurons in terms of system efficiency and high performance.
27 May 2021 in IEEE Transactions on Electron Devices
In this work, a novel synaptic array architecture enabling forward propagation (FP) and backward propagation (BP) in the nand flash memory is proposed for the first time for on-chip learning. In the proposed synaptic architecture, positive synaptic weight and negative synaptic weight are separated in different arrays to enable weights to be transposed correctly. In addition, source-lines (SLs) are separated, which is different from the conventional nand flash memory, to enable both the FP and BP in the nand flash memory. By applying input and error input to bitlines (BLs) and string-select lines (SSLs) in nand cell array, respectively, accurate vector-matrix multiplication is successfully carried out in both FP and BP eliminating the effect of pass cells. At a read voltage of 2 V, the inference accuracy of 95.58% which is comparable to that of 95.81% obtained with perfect linear device is achieved. The proposed on-chip learning system is much more robust to weight variation compared to the off-chip learning system. Finally, superiority of the proposed on-chip learning architecture is verified by circuit simulation of a neural network.
An oxide fin-based AND flash memory synaptic device is proposed and fabricated using a spacer patterning technology for a hardware-based binary neural network (BNN). A fin-like curved channel structure provides local electric field enhancement, which improves programming efficiency compared to planar-type flash synaptic devices. The fin-based AND flash cell exhibits a high on/off current ratio (>10 5 ) with sub-pA off current, and a low programming voltage (< 9 V) is used to achieve a sufficient dynamic range of synaptic weights (>10 3 ) for BNNs. Furthermore, a hardware-based BNN using novel two cell-based synaptic devices arranged in AND array architecture is proposed to implement parallel XNOR operation and bit-counting. Proposed BNN using the synapse model with measured dynamic range and retention property shows only < 0.5 % degradation of classification accuracy compared to the baseline accuracy, which is suitable to perform off-chip event-driven computation using parallel read-out operations.