Spiking Neural Network Model Research Articles

A Spiking Neural Network with Adaptive Graph Convolution and LSTM for EEG-Based Brain-Computer Interfaces.

Electroencephalography (EEG) signals classification is essential for the brain-computer interface (BCI). Recently, energy-efficient spiking neural networks (SNNs) have shown great potential in EEG analysis due to their ability to capture the complex dynamic properties of biological neurons while also processing stimulus information through precisely timed spike trains. However, most existing methods do not effectively mine the specific spatial topology of EEG channels and temporal dependencies of the encoded EEG spikes. Moreover, most are designed for specific BCI tasks and lack some generality. Hence, this study presents a novel SNN model with the customized spike-based adaptive graph convolution and long short-term memory (LSTM), termed SGLNet, for EEG-based BCIs. Specifically, we first adopt a learnable spike encoder to convert the raw EEG signals into spike trains. Then, we tailor the concepts of the multi-head adaptive graph convolution to SNN so that it can make good use of the intrinsic spatial topology information among distinct EEG channels. Finally, we design the spike-based LSTM units to further capture the temporal dependencies of the spikes. We evaluate our proposed model on two publicly available datasets from two representative fields of BCI, notably emotion recognition, and motor imagery decoding. The empirical evaluations demonstrate that SGLNet consistently outperforms existing state-of-the-art EEG classification algorithms. This work provides a new perspective for exploring high-performance SNNs for future BCIs with rich spatiotemporal dynamics.

STSC-SNN: Spatio-Temporal Synaptic Connection with temporal convolution and attention for spiking neural networks.

Spiking neural networks (SNNs), as one of the algorithmic models in neuromorphic computing, have gained a great deal of research attention owing to temporal information processing capability, low power consumption, and high biological plausibility. The potential to efficiently extract spatio-temporal features makes it suitable for processing event streams. However, existing synaptic structures in SNNs are almost full-connections or spatial 2D convolution, neither of which can extract temporal dependencies adequately. In this work, we take inspiration from biological synapses and propose a Spatio-Temporal Synaptic Connection SNN (STSC-SNN) model to enhance the spatio-temporal receptive fields of synaptic connections, thereby establishing temporal dependencies across layers. Specifically, we incorporate temporal convolution and attention mechanisms to implement synaptic filtering and gating functions. We show that endowing synaptic models with temporal dependencies can improve the performance of SNNs on classification tasks. In addition, we investigate the impact of performance via varied spatial-temporal receptive fields and reevaluate the temporal modules in SNNs. Our approach is tested on neuromorphic datasets, including DVS128 Gesture (gesture recognition), N-MNIST, CIFAR10-DVS (image classification), and SHD (speech digit recognition). The results show that the proposed model outperforms the state-of-the-art accuracy on nearly all datasets.

Adversarial attacks on spiking convolutional neural networks for event-based vision.

Event-based dynamic vision sensors provide very sparse output in the form of spikes, which makes them suitable for low-power applications. Convolutional spiking neural networks model such event-based data and develop their full energy-saving potential when deployed on asynchronous neuromorphic hardware. Event-based vision being a nascent field, the sensitivity of spiking neural networks to potentially malicious adversarial attacks has received little attention so far. We show how white-box adversarial attack algorithms can be adapted to the discrete and sparse nature of event-based visual data, and demonstrate smaller perturbation magnitudes at higher success rates than the current state-of-the-art algorithms. For the first time, we also verify the effectiveness of these perturbations directly on neuromorphic hardware. Finally, we discuss the properties of the resulting perturbations, the effect of adversarial training as a defense strategy, and future directions.

Bio-inspired Active Learning method in spiking neural network

Spiking neural networks (SNNs) have gained a lot of attention and achievements recently because of their low-power advantages on neuromorphic hardware. However, training deep SNNs still requires a large number of labeled data which are expensive to obtain. To address this issue, we propose an effective Bio-inspired Active Learning (BAL) method in this paper to reduce the training cost of SNN models. Specifically, bio-inspired behavior patterns of spiking neurons are defined to represent the internal states of SNN models for active learning. Then, an active learning sample selection strategy is proposed by leveraging the empirical and generalization pattern divergence in SNNs. By labeling selected samples and adding them to training, behavioral patterns can be optimized to improve the performance of neural networks. Comprehensive experiments are conducted on the CIFAR-10, SVHN, and Fashion-MNIST datasets with various sample proportions. The experimental results demonstrate that the proposed BAL achieves state-of-the-arts performance in SNNs compared with the existing active learning methods.

Fault Detection in Electrical Equipment by Infrared Thermography Images Using Spiking Neural Network Through Hybrid Feature Selection

The Spiking Neural Network (SNN) model is a third-generation neural network model that uses spike pulse trains to detect and classify electrical problems in the electrical equipment under consideration. To diagnose the thermal issue in the early stages, it is required to evaluate and monitor the electrical components. The detection method in Infrared Thermography (IRT) is hybrid optimization, which is a Dragonfly Algorithm (DA)–Ant Colony Optimization (ACO) technique that produces a higher exploration and exploitation rate while moving toward an optimal solution with a higher convergence rate. The innovative hybrid DA–ACO method used in this study aims to find the best weights for the SNN model while also extracting the most useful characteristics for defect detection in electrical equipment. To improve classification accuracy, the optimum features picked using the created hybrid DA–ACO are provided to the SNN model. Based on conventional and other optimization methods, the proposed method provided superior results from the execution results.

Creation of Neuronal Ensembles and Cell-Specific Homeostatic Plasticity through Chronic Sparse Optogenetic Stimulation.

Cortical computations emerge from the dynamics of neurons embedded in complex cortical circuits. Within these circuits, neuronal ensembles, which represent subnetworks with shared functional connectivity, emerge in an experience-dependent manner. Here we induced ensembles in ex vivo cortical circuits from mice of either sex by differentially activating subpopulations through chronic optogenetic stimulation. We observed a decrease in voltage correlation, and importantly a synaptic decoupling between the stimulated and nonstimulated populations. We also observed a decrease in firing rate during Up-states in the stimulated population. These ensemble-specific changes were accompanied by decreases in intrinsic excitability in the stimulated population, and a decrease in connectivity between stimulated and nonstimulated pyramidal neurons. By incorporating the empirically observed changes in intrinsic excitability and connectivity into a spiking neural network model, we were able to demonstrate that changes in both intrinsic excitability and connectivity accounted for the decreased firing rate, but only changes in connectivity accounted for the observed decorrelation. Our findings help ascertain the mechanisms underlying the ability of chronic patterned stimulation to create ensembles within cortical circuits and, importantly, show that while Up-states are a global network-wide phenomenon, functionally distinct ensembles can preserve their identity during Up-states through differential firing rates and correlations.SIGNIFICANCE STATEMENT The connectivity and activity patterns of local cortical circuits are shaped by experience. This experience-dependent reorganization of cortical circuits is driven by complex interactions between different local learning rules, external input, and reciprocal feedback between many distinct brain areas. Here we used an ex vivo approach to demonstrate how simple forms of chronic external stimulation can shape local cortical circuits in terms of their correlated activity and functional connectivity. The absence of feedback between different brain areas and full control of external input allowed for a tractable system to study the underlying mechanisms and development of a computational model. Results show that differential stimulation of subpopulations of neurons significantly reshapes cortical circuits and forms subnetworks referred to as neuronal ensembles.

Production of adaptive movement patterns via an insect inspired spiking neural network central pattern generator.

Navigation in ever-changing environments requires effective motor behaviors. Many insects have developed adaptive movement patterns which increase their success in achieving navigational goals. A conserved brain area in the insect brain, the Lateral Accessory Lobe, is involved in generating small scale search movements which increase the efficacy of sensory sampling. When the reliability of an essential navigational stimulus is low, searching movements are initiated whereas if the stimulus reliability is high, a targeted steering response is elicited. Thus, the network mediates an adaptive switching between motor patterns. We developed Spiking Neural Network models to explore how an insect inspired architecture could generate adaptive movements in relation to changing sensory inputs. The models are able to generate a variety of adaptive movement patterns, the majority of which are of the zig-zagging kind, as seen in a variety of insects. Furthermore, these networks are robust to noise. Because a large spread of network parameters lead to the correct movement dynamics, we conclude that the investigated network architecture is inherently well-suited to generating adaptive movement patterns.

Dynamic control of eye-head gaze shifts by a spiking neural network model of the superior colliculus.

To reorient gaze (the eye's direction in space) towards a target is an overdetermined problem, as infinitely many combinations of eye- and head movements can specify the same gaze-displacement vector. Yet, behavioral measurements show that the primate gaze-control system selects a specific contribution of eye- and head movements to the saccade, which depends on the initial eye-in-head orientation. Single-unit recordings in the primate superior colliculus (SC) during head-unrestrained gaze shifts have further suggested that cells may encode the instantaneous trajectory of a desired straight gaze path in a feedforward way by the total cumulative number of spikes in the neural population, and that the instantaneous gaze kinematics are thus determined by the neural firing rates. The recordings also indicated that the latter is modulated by the initial eye position. We recently proposed a conceptual model that accounts for many of the observed properties of eye-head gaze shifts and on the potential role of the SC in gaze control. Here, we extend and test the model by incorporating a spiking neural network of the SC motor map, the output of which drives the eye-head motor control circuitry by linear cumulative summation of individual spike effects of each recruited SC neuron. We propose a simple neural mechanism on SC cells that explains the modulatory influence of feedback from an initial eye-in-head position signal on their spiking activity. The same signal also determines the onset delay of the head movement with respect to the eye. Moreover, the downstream eye- and head burst generators were taken to be linear, as our earlier work had indicated that much of the non-linear main-sequence kinematics of saccadic eye movements may be due to neural encoding at the collicular level, rather than at the brainstem. We investigate how the spiking activity of the SC population drives gaze to the intended target location within a dynamic local gaze-velocity feedback circuit that yields realistic eye- and head-movement kinematics and dynamic SC gaze-movement fields.

Extended liquid state machines for speech recognition.

A liquid state machine (LSM) is a biologically plausible model of a cortical microcircuit. It exists of a random, sparse reservoir of recurrently connected spiking neurons with fixed synapses and a trainable readout layer. The LSM exhibits low training complexity and enables backpropagation-free learning in a powerful, yet simple computing paradigm. In this work, the liquid state machine is enhanced by a set of bio-inspired extensions to create the extended liquid state machine (ELSM), which is evaluated on a set of speech data sets. Firstly, we ensure excitatory/inhibitory (E/I) balance to enable the LSM to operate in edge-of-chaos regime. Secondly, spike-frequency adaptation (SFA) is introduced in the LSM to improve the memory capabilities. Lastly, neuronal heterogeneity, by means of a differentiation in time constants, is introduced to extract a richer dynamical LSM response. By including E/I balance, SFA, and neuronal heterogeneity, we show that the ELSM consistently improves upon the LSM while retaining the benefits of the straightforward LSM structure and training procedure. The proposed extensions led up to an 5.2% increase in accuracy while decreasing the number of spikes in the ELSM up to 20.2% on benchmark speech data sets. On some benchmarks, the ELSM can even attain similar performances as the current state-of-the-art in spiking neural networks. Furthermore, we illustrate that the ELSM input-liquid and recurrent synaptic weights can be reduced to 4-bit resolution without any significant loss in classification performance. We thus show that the ELSM is a powerful, biologically plausible and hardware-friendly spiking neural network model that can attain near state-of-the-art accuracy on speech recognition benchmarks for spiking neural networks.

Neuromodulator-dependent synaptic tagging and capture retroactively controls neural coding in spiking neural networks

Events that are important to an individual’s life trigger neuromodulator release in brain areas responsible for cognitive and behavioral function. While it is well known that the presence of neuromodulators such as dopamine and norepinephrine is required for memory consolidation, the impact of neuromodulator concentration is, however, less understood. In a recurrent spiking neural network model featuring neuromodulator-dependent synaptic tagging and capture, we study how synaptic memory consolidation depends on the amount of neuromodulator present in the minutes to hours after learning. We find that the storage of rate-based and spike timing-based information is controlled by the level of neuromodulation. Specifically, we find better recall of temporal information for high levels of neuromodulation, while we find better recall of rate-coded spatial patterns for lower neuromodulation, mediated by the selection of different groups of synapses for consolidation. Hence, our results indicate that in minutes to hours after learning, the level of neuromodulation may alter the process of synaptic consolidation to ultimately control which type of information becomes consolidated in the recurrent neural network.

Static hand gesture recognition for American sign language using neuromorphic hardware

In this paper, we develop four spiking neural network (SNN) models for two static American sign language (ASL) hand gesture classification tasks, i.e., the ASL alphabet and ASL digits. The SNN models are deployed on Intel’s neuromorphic platform, Loihi, and then compared against equivalent deep neural network (DNN) models deployed on an edge computing device, the Intel neural compute stick 2 (NCS2). We perform a comprehensive comparison between the two systems in terms of accuracy, latency, power consumption, and energy. The best DNN model achieves an accuracy of 99.93% on the ASL alphabet dataset, whereas the best performing SNN model has an accuracy of 99.30%. For the ASL-digits dataset, the best DNN model achieves an accuracy of 99.76% accuracy while the SNN achieves 99.03%. Moreover, our obtained experimental results show that the Loihi neuromorphic hardware implementations achieve up to 20.64× and 4.10× reduction in power consumption and energy, respectively, when compared to NCS2.

Modular building blocks for mapping spiking neural networks onto a programmable neuromorphic processor

In the past few years, brain-inspired artificial intelligence (AI) system with energy-efficient neuromorphic computing has gathered lots of interests. However, how to build an efficient algorithm-to-hardware development tool for spiking neural networks (SNNs) remains a big challenge. In this paper, we present a novel design of neural controller which can flexibly generate the store-and-release signals in the designed SNNs. Further, we propose a mapping methodology with modular building blocks to deploy SNN models onto a many-core programmable neuromorphic processor. Experimental results show that the presented mapping method can be adaptive to various SNN layers of different sizes and quantization precisions. Besides, our demonstrated system on chip can achieve about 181 and 26 images per second runtime inference speed on MNIST and CIFAR-10 dataset respectively and show comparable accuracies and significantly better power performance with their artificial neural network (ANN) counterparts running on traditional GPU.

Simulations of working memory spiking networks driven by short-term plasticity.

Working Memory (WM) is a cognitive mechanism that enables temporary holding and manipulation of information in the human brain. This mechanism is mainly characterized by a neuronal activity during which neuron populations are able to maintain an enhanced spiking activity after being triggered by a short external cue. In this study, we implement, using the NEST simulator, a spiking neural network model in which the WM activity is sustained by a mechanism of short-term synaptic facilitation related to presynaptic calcium kinetics. The model, which is characterized by leaky integrate-and-fire neurons with exponential postsynaptic currents, is able to autonomously show an activity regime in which the memory information can be stored in a synaptic form as a result of synaptic facilitation, with spiking activity functional to facilitation maintenance. The network is able to simultaneously keep multiple memories by showing an alternated synchronous activity which preserves the synaptic facilitation within the neuron populations holding memory information. The results shown in this study confirm that a WM mechanism can be sustained by synaptic facilitation.

Temporal Coding in Spiking Neural Networks With Alpha Synaptic Function: Learning With Backpropagation.

The timing of individual neuronal spikes is essential for biological brains to make fast responses to sensory stimuli. However, conventional artificial neural networks lack the intrinsic temporal coding ability present in biological networks. We propose a spiking neural network model that encodes information in the relative timing of individual spikes. In classification tasks, the output of the network is indicated by the first neuron to spike in the output layer. This temporal coding scheme allows the supervised training of the network with backpropagation, using locally exact derivatives of the postsynaptic spike times with respect to presynaptic spike times. The network operates using a biologically plausible synaptic transfer function. In addition, we use trainable pulses that provide bias, add flexibility during training, and exploit the decayed part of the synaptic function. We show that such networks can be successfully trained on multiple data sets encoded in time, including MNIST. Our model outperforms comparable spiking models on MNIST and achieves similar quality to fully connected conventional networks with the same architecture. The spiking network spontaneously discovers two operating modes, mirroring the accuracy-speed tradeoff observed in human decision-making: a highly accurate but slow regime, and a fast but slightly lower accuracy regime. These results demonstrate the computational power of spiking networks with biological characteristics that encode information in the timing of individual neurons. By studying temporal coding in spiking networks, we aim to create building blocks toward energy-efficient, state-based biologically inspired neural architectures. We provide open-source code for the model.

A Deep Spiking Neural Network Anomaly Detection Method.

Cyber-attacks on specialized industrial control systems are increasing in frequency and sophistication, which means stronger countermeasures need to be implemented, requiring the designers of the equipment in question to re-evaluate and redefine their methods for actively protecting against advanced mass cyber-attacks. The attacks in question have huge motivations, ranging from corporate espionage to political targets, but in any case, they have a substantial financial impact and severe real-world implications. It should also be said that it is challenging to defend against cyber threats because a single point of entry can be enough to destroy an entire organization or put it out of business. This paper examines threats to the digital security of vibration monitoring systems used in petroleum infrastructure protection services, such as pipelines, pumps, and tank farms, where malicious interventions can cause explosions, fires, or toxic releases, with incalculable economic and environmental consequences. Specifically, a deep spiking neural network anomaly detection method is presented, which models the spike sequences and the internal presentation mechanisms of the information to discover with very high accuracy anomalies in vibration analysis systems used in oil infrastructure protection services. This is achieved by simulating the complex structures of the human brain and the way neural information is processed and transmitted. This work uses a particularly innovative form of the Galves–Löcherbach Spiking Model (GLSM) [1], which is a spiking neural network model with intrinsic stochasticity, ideal for modeling complex spatiotemporal situations, which is enhanced with possibilities of exploiting confidence intervals by modeling optimally stochastic variable-length memory chains that have a finite state space.

MAP-SNN: Mapping spike activities with multiplicity, adaptability, and plasticity into bio-plausible spiking neural networks.

Spiking Neural Networks (SNNs) are considered more biologically realistic and power-efficient as they imitate the fundamental mechanism of the human brain. Backpropagation (BP) based SNN learning algorithms that utilize deep learning frameworks have achieved good performance. However, those BP-based algorithms partially ignore bio-interpretability. In modeling spike activity for biological plausible BP-based SNNs, we examine three properties: multiplicity, adaptability, and plasticity (MAP). Regarding multiplicity, we propose a Multiple-Spike Pattern (MSP) with multiple-spike transmission to improve model robustness in discrete time iterations. To realize adaptability, we adopt Spike Frequency Adaption (SFA) under MSP to reduce spike activities for enhanced efficiency. For plasticity, we propose a trainable state-free synapse that models spike response current to increase the diversity of spiking neurons for temporal feature extraction. The proposed SNN model achieves competitive performances on the N-MNIST and SHD neuromorphic datasets. In addition, experimental results demonstrate that the proposed three aspects are significant to iterative robustness, spike efficiency, and the capacity to extract spikes' temporal features. In summary, this study presents a realistic approach for bio-inspired spike activity with MAP, presenting a novel neuromorphic perspective for incorporating biological properties into spiking neural networks.

Neuromorphic circuit based on the un-supervised learning of biologically inspired spiking neural network for pattern recognition

One of the most sophisticated platforms for hosting intelligent systems is bio-inspired. This study proposes pattern recognition hardware using a biologically inspired Spiking Neural Network (SNN) and the new dimensionality reduction approach. The SNN model is based on real neural networks consisting of spiking neurons linked by excitatory and inhibitory synapses activated by excitatory and inhibitory neurotransmitters. Also, a semi-supervised (un-supervised STDP based learning with supervised weight initialization), spike-based learning strategy based on the learning procedure of the nervous system is used to teach the spiking output layer neurons, albeit that the hardware implementation benefits from a semi-supervised approach. The goal of this research is to accurately categorize patterns in the MNIST and CIFAR10 datasets using an SNN-based hardware platform. Due to the limitations of the latter’s resources, a dimensionality reduction based on principal component analysis (PCA) is proposed to speed up the processing procedure and reduce the hardware implementation cost. The presented pattern recognition platform is implemented using the Xilinx® VIVADO high-level synthesis platform (HLS). Finally, optimization approaches are used to improve the used space, reduce hardware implementation delay, and speed up the design process.

High-efficiency and low-energy ship recognition strategy based on spiking neural network in SAR images.

Ship recognition using synthetic aperture radar (SAR) images has important applications in the military and civilian fields. Aiming at the problems of the many model parameters and high-energy losses in the traditional deep learning methods for the target recognition in the SAR images, this study has proposed a high-efficiency and low-energy ship recognition strategy based on the spiking neural network (SNN) in the SAR images. First, the visual attention mechanism is used to extract the visual saliency map from the SAR image, and then the Poisson encoder is used to encode it into a spike train, which can suppress the background noise while retaining the visual saliency feature of the SAR image. Besides, an SNN model integrating the time-series information is constructed by combining the leaked and integrated firing spiking neurons with the convolutional neural network (CNN), which can use the firing frequency of the spiking neurons to realize the ship recognition in SAR images. Finally, to solve the problem that SNN model is difficult to train, the arctangent function is used as the surrogate gradient function of the spike emission function during the backpropagation. Hence, applying this backpropagation method to the training process can optimize the SNN model. The experimental results show the following: (1) the proposed strategy can more accurately recognize the ship in the SAR image, and the F1 score can reach 98.55%, which has a better recognition performance than the other traditional deep learning methods; (2) the proposed strategy has the least amount of model parameters (only 3.11MB), which is far less than the model parameters of the other traditional deep learning methods; (3) the proposed strategy has fewer operations (only 17.97M) and can reach 1/30 time of operands of the other traditional deep learning methods, which shows the high efficiency of the proposed strategy using the spike emission signals; (4) the proposed strategy has the energy loss of 1.38 × 10−6J, which can achieve the low energy advantage of nearly three orders of the magnitude compared to the other traditional deep learning methods, indicating that the proposed strategy has a significant energy efficiency.

Personalized Spiking Neural Network Models of Clinical and Environmental Factors to Predict Stroke

The high incidence of stroke occurrence necessitates the understanding of its causes and possible ways for early prediction and prevention. In this respect, statistical methods offer the “big picture,” but they have a weak predictive ability at an individual level. This research proposes a new personalized modeling method based on computational spiking neural networks (SNN) for the identification of causal associations between clinical and environmental time series data that can be used to predict individual stroke events. The method is tested on 804 stroke patients. Given a clinical data set of patients who experienced a stroke in the past and the corresponding environmental time-series data for a selected time-window before the stroke event, the method identifies the clusters of individuals with a high risk for stroke under similar conditions. The methodology involves a pipeline of processes when creating a personalized model for an individual x: (1) selecting a group of individuals Gx with similar personal records to x; (2) training a personalized SNN x model of several days of environmental data related to the Gx group to predict the risk of stroke for x at least one day earlier; (3) model interpretability through 3D visualization; (4) discovery of personalized predictive markers. The results are twofold, first proposing a new computational methodology and second presenting new findings. It is found that certain environmental factors, such as SO2, PM10, CO, and PM2.5, increase the risk of stroke if an individual x belongs to a certain cluster of people, characterized by a combination of family history of stroke and diabetes, overweight, vascular/heart disease, age, and other. For the used population data, the proposed method can predict accurately individual risk of stroke before the day of the stroke. The paper presents a new methodology for personalized machine learning methods to define subgroups of the population with a high risk of stroke and to predict early individual risk of the stroke event. This makes the proposed cognitive computation method useful to reduce morbidity and mortality in society. The method is broadly applicable for predicting individual risk of other diseases and mental health conditions.

EnforceSNN: Enabling resilient and energy-efficient spiking neural network inference considering approximate DRAMs for embedded systems.

Spiking Neural Networks (SNNs) have shown capabilities of achieving high accuracy under unsupervised settings and low operational power/energy due to their bio-plausible computations. Previous studies identified that DRAM-based off-chip memory accesses dominate the energy consumption of SNN processing. However, state-of-the-art works do not optimize the DRAM energy-per-access, thereby hindering the SNN-based systems from achieving further energy efficiency gains. To substantially reduce the DRAM energy-per-access, an effective solution is to decrease the DRAM supply voltage, but it may lead to errors in DRAM cells (i.e., so-called approximate DRAM). Toward this, we propose EnforceSNN, a novel design framework that provides a solution for resilient and energy-efficient SNN inference using reduced-voltage DRAM for embedded systems. The key mechanisms of our EnforceSNN are: (1) employing quantized weights to reduce the DRAM access energy; (2) devising an efficient DRAM mapping policy to minimize the DRAM energy-per-access; (3) analyzing the SNN error tolerance to understand its accuracy profile considering different bit error rate (BER) values; (4) leveraging the information for developing an efficient fault-aware training (FAT) that considers different BER values and bit error locations in DRAM to improve the SNN error tolerance; and (5) developing an algorithm to select the SNN model that offers good trade-offs among accuracy, memory, and energy consumption. The experimental results show that our EnforceSNN maintains the accuracy (i.e., no accuracy loss for BER ≤ 10−3) as compared to the baseline SNN with accurate DRAM while achieving up to 84.9% of DRAM energy saving and up to 4.1x speed-up of DRAM data throughput across different network sizes.