14 February 2023 in IEEE Journal on Emerging and Selected Topics in Circuits and Systems
A novel spiking neuromorphic architecture is presented in this paper. The architecture is based on charge-trap transistors (CTTs) which are experimentally-verified compute-in-memory devices. The proposed low-power scalable architecture targets large neural network applications, such as machine learning tasks and emulation of brain connectivity networks. Data within the proposed architecture is encoded using a number of spikes approach. The CTT-based synapses receive Gaussian spikes, the most energy-efficient waveform for communication, as inputs from other neurons, the spikes are multiplied by synaptic weights and accumulated. The neuron, designed using a leaky integrate and fire model, generates a similar spike at the output. The proposed architecture is compared to literature and exhibits superior parameters. The neuron (including the synaptic array) occupies an area of 178.25 μm 2 , supporting 5.6k neurons and 560k synapses per mm 2 , as well as exhibits low energy per synaptic operation of 8 pJ. To validate the proposed architecture, a single neuron was designed and evaluated as a binary classifier for two numbers from the MNIST data set. The accuracy, recall, and precision of the hardware neuron for the binary classification task are, respectively, 99.2%, 99.5%, and 98.6% (similar to results from other reported works).
5 March 2026 in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
Neuromorphic hardware combined with spiking neural networks (SNNs) supports low-power, highly parallel AI inference. Inference accuracy can drop sharply when models trained under ideal software assumptions are deployed on inherently noisy hardware. This noise is especially evident in the current of subthreshold operated charge-trap transistors (CTTs) representing network weights. Although CTTs are CMOS-compatible and attractive for large-scale neuromorphic systems due to their energy and cost efficiency, programming imprecision and process, voltage, and temperature (PVT) variations can significantly alter their current, degrading network accuracy. In this article, CTT weight and neuron sources of variability are first identified. Multiparameter noise models are then derived from the hardware components, and sampled noise is injected into both weights and neuron parameters during training, improving robustness across a wide range of noise levels without increasing model size or significantly increasing training time. A novel training algorithm is introduced in which each mini-batch is evaluated multiple times under different noise samples, with the loss optimized to enforce consistency across various noise values. The software behavioral model is verified using Cadence simulations across 45 binary classifiers. The accuracy of ideal-trained classifiers degrades by up to 44.2% under hardware-induced noise, while the proposed training algorithm improves accuracy by up to 23.34%. The validated framework and training algorithm are then used for larger networks performing image classification on CIFAR-10 and CIFAR-100 datasets, improving accuracy from a random state under noise to competitive levels relative to state-of-the-art, with the main overhead being additional epochs determined by noise range rather than network size.