Event-Driven Continuous STDP Learning With Deep Structure for Visual Pattern Recognition

Deep Spike Learning With Local Classifiers

Backpropagation has been successfully generalized to optimize deep spiking neural networks (SNNs), where, nevertheless, gradients need to be propagated back through all layers, resulting in a massive consumption of computing resources and an obstacle to the parallelization of training. A biologically motivated scheme of local learning provides an alternative to efficiently train deep networks but often suffers a low performance of accuracy on practical tasks. Thus, how to train deep SNNs with the local learning scheme to achieve both efficient and accurate performance still remains an important challenge. In this study, we focus on a supervised local learning scheme where each layer is independently optimized with an auxiliary classifier. Accordingly, we first propose a spike-based efficient local learning rule by only considering the direct dependencies in the current time. We then propose two variants that additionally incorporate temporal dependencies through a backward and forward process, respectively. The effectiveness and performance of our proposed methods are extensively evaluated with six mainstream datasets. Experimental results show that our methods can successfully scale up to large networks and substantially outperform the spike-based local learning baselines on all studied benchmarks. Our results also reveal that gradients with temporal dependencies are essential for high performance on temporal tasks, while they have negligible effects on rate-based tasks. Our work is significant as it brings the performance of spike-based local learning to a new level with the computational benefits being retained.

A Time-Delay Feedback Neural Network for Discriminating Small, Fast-Moving Targets in Complex Dynamic Environments

Discriminating small moving objects within complex visual environments is a significant challenge for autonomous micro-robots that are generally limited in computational power. By exploiting their highly evolved visual systems, flying insects can effectively detect mates and track prey during rapid pursuits, even though the small targets equate to only a few pixels in their visual field. The high degree of sensitivity to small target movement is supported by a class of specialized neurons called small target motion detectors (STMDs). Existing STMD-based computational models normally comprise four sequentially arranged neural layers interconnected via feedforward loops to extract information on small target motion from raw visual inputs. However, feedback, another important regulatory circuit for motion perception, has not been investigated in the STMD pathway and its functional roles for small target motion detection are not clear. In this article, we propose an STMD-based neural network with feedback connection (feedback STMD), where the network output is temporally delayed, then fed back to the lower layers to mediate neural responses. We compare the properties of the model with and without the time-delay feedback loop and find that it shows a preference for high-velocity objects. Extensive experiments suggest that the feedback STMD achieves superior detection performance for fast-moving small targets, while significantly suppressing background false positive movements which display lower velocities. The proposed feedback model provides an effective solution in robotic visual systems for detecting fast-moving small targets that are always salient and potentially threatening.

Complementary Memtransistor-Based Multilayer Neural Networks for Online Supervised Learning Through (Anti-)Spike-Timing-Dependent Plasticity

We propose a complete hardware-based architecture of multilayer neural networks (MNNs), including electronic synapses, neurons, and periphery circuitry to implement supervised learning (SL) algorithm of extended remote supervised method (ReSuMe). In this system, complementary (a pair of n- and p-type) memtransistors (C-MTs) are used as an electrical synapse. By applying the learning rule of spike-timing-dependent plasticity (STDP) to the memtransistor connecting presynaptic neuron to the output one whereas the contrary anti-STDP rule to the other memtransistor connecting presynaptic neuron to the teacher one, extended ReSuMe with multiple layers is realized without the usage of those complicated supervising modules in previous approaches. In this way, both the C-MT-based chip area and power consumption of the learning circuit for weight updating operation are drastically decreased comparing with the conventional single memtransistor (S-MT)-based designs. Two typical benchmarks, the linearly nonseparable benchmark XOR problem and Mixed National Institute of Standards and Technology database (MNIST) recognition have been successfully tackled using the proposed MNN system while impact of the nonideal factors of realistic devices has been evaluated.

Toward the Optimal Design and FPGA Implementation of Spiking Neural Networks

The performance of a biologically plausible spiking neural network (SNN) largely depends on the model parameters and neural dynamics. This article proposes a parameter optimization scheme for improving the performance of a biologically plausible SNN and a parallel on-field-programmable gate array (FPGA) online learning neuromorphic platform for the digital implementation based on two numerical methods, namely, the Euler and third-order Runge-Kutta (RK3) methods. The optimization scheme explores the impact of biological time constants on information transmission in the SNN and improves the convergence rate of the SNN on digit recognition with a suitable choice of the time constants. The parallel digital implementation leads to a significant speedup over software simulation on a general-purpose CPU. The parallel implementation with the Euler method enables around 180× ( 20× ) training (inference) speedup over a Pytorch-based SNN simulation on CPU. Moreover, compared with previous work, our parallel implementation shows more than 300× ( 240× ) improvement on speed and 180× ( 250× ) reduction in energy consumption for training (inference). In addition, due to the high-order accuracy, the RK3 method is demonstrated to gain 2× training speedup over the Euler method, which makes it suitable for online training in real-time applications.

SAM: A Unified Self-Adaptive Multicompartmental Spiking Neuron Model for Learning With Working Memory

Working memory is a fundamental feature of biological brains for perception, cognition, and learning. In addition, learning with working memory, which has been show in conventional artificial intelligence systems through recurrent neural networks, is instrumental to advanced cognitive intelligence. However, it is hard to endow a simple neuron model with working memory, and to understand the biological mechanisms that have resulted in such a powerful ability at the neuronal level. This article presents a novel self-adaptive multicompartment spiking neuron model, referred to as SAM, for spike-based learning with working memory. SAM integrates four major biological principles including sparse coding, dendritic non-linearity, intrinsic self-adaptive dynamics, and spike-driven learning. We first describe SAM's design and explore the impacts of critical parameters on its biological dynamics. We then use SAM to build spiking networks to accomplish several different tasks including supervised learning of the MNIST dataset using sequential spatiotemporal encoding, noisy spike pattern classification, sparse coding during pattern classification, spatiotemporal feature detection, meta-learning with working memory applied to a navigation task and the MNIST classification task, and working memory for spatiotemporal learning. Our experimental results highlight the energy efficiency and robustness of SAM in these wide range of challenging tasks. The effects of SAM model variations on its working memory are also explored, hoping to offer insight into the biological mechanisms underlying working memory in the brain. The SAM model is the first attempt to integrate the capabilities of spike-driven learning and working memory in a unified single neuron with multiple timescale dynamics. The competitive performance of SAM could potentially contribute to the development of efficient adaptive neuromorphic computing systems for various applications from robotics to edge computing.

A Hybrid Recursive Implementation of Broad Learning With Incremental Features

The broad learning system (BLS) paradigm has recently emerged as a computationally efficient approach to supervised learning. Its efficiency arises from a learning mechanism based on the method of least-squares. However, the need for storing and inverting large matrices can put the efficiency of such mechanism at risk in big-data scenarios. In this work, we propose a new implementation of BLS in which the need for storing and inverting large matrices is avoided. The distinguishing features of the designed learning mechanism are as follows: 1) the training process can balance between efficient usage of memory and required iterations (hybrid recursive learning) and 2) retraining is avoided when the network is expanded (incremental learning). It is shown that, while the proposed framework is equivalent to the standard BLS in terms of trained network weights,much larger networks than the standard BLS can be smoothly trained by the proposed solution, projecting BLS toward the big-data frontier.

Toward Efficient Processing and Learning With Spikes: New Approaches for Multispike Learning

Spikes are the currency in central nervous systems for information transmission and processing. They are also believed to play an essential role in low-power consumption of the biological systems, whose efficiency attracts increasing attentions to the field of neuromorphic computing. However, efficient processing and learning of discrete spikes still remain a challenging problem. In this article, we make our contributions toward this direction. A simplified spiking neuron model is first introduced with the effects of both synaptic input and firing output on the membrane potential being modeled with an impulse function. An event-driven scheme is then presented to further improve the processing efficiency. Based on the neuron model, we propose two new multispike learning rules which demonstrate better performance over other baselines on various tasks, including association, classification, and feature detection. In addition to efficiency, our learning rules demonstrate high robustness against the strong noise of different types. They can also be generalized to different spike coding schemes for the classification task, and notably, the single neuron is capable of solving multicategory classifications with our learning rules. In the feature detection task, we re-examine the ability of unsupervised spike-timing-dependent plasticity with its limitations being presented, and find a new phenomenon of losing selectivity. In contrast, our proposed learning rules can reliably solve the task over a wide range of conditions without specific constraints being applied. Moreover, our rules cannot only detect features but also discriminate them. The improved performance of our methods would contribute to neuromorphic computing as a preferable choice.

Federated Continuous Learning With Broad Network Architecture

Federated learning (FL) is a machine-learning setting, where multiple clients collaboratively train a model under the coordination of a central server. The clients' raw data are locally stored, and each client only uploads the trained weight to the server, which can mitigate the privacy risks from the centralized machine learning. However, most of the existing FL models focus on one-time learning without consideration for continuous learning. Continuous learning supports learning from streaming data continuously, so it can adapt to environmental changes and provide better real-time performance. In this article, we present a federated continuous learning scheme based on broad learning (FCL-BL) to support efficient and accurate federated continuous learning (FCL). In FCL-BL, we propose a weighted processing strategy to solve the catastrophic forgetting problem, so FCL-BL can handle continuous learning. Then, we develop a local-independent training solution to support fast and accurate training in FCL-BL. The proposed solution enables us to avoid using a time-consuming synchronous approach while addressing the inaccurate-training issue rooted in the previous asynchronous approach. Moreover, we introduce a batch-asynchronous approach and broad learning (BL) technique to guarantee the high efficiency of FCL-BL. Specifically, the batch-asynchronous approach reduces the number of client-server interaction rounds, and the BL technique supports incremental learning without retraining when learning newly produced data. Finally, theoretical analysis and experimental results further illustrate that FCL-BL is superior to the existing FL schemes in terms of efficiency and accuracy in FCL.

Pattern Recognition of Grating Perimeter Intrusion Behavior in Deep Learning Method

An intrusion behavior recognition method based on deep learning is proposed in this paper in order to improve the recognition accuracy of raster perimeter intrusion behavior. The Mach–Zehnder fiber optic interferometer was used to collect the external vibration signal sensing unit, capture the external vibration signal, use the cross-correlation characteristic method to obtain the minimum frame length of the fiber vibration signal, and preprocess the intrusion signal according to the signal strength. The intrusion signals were superimposed and several sections of signals were intercepted by fixed window length; the spectrum information is obtained by Fourier transform of the intercepted stationary signals. The convolution neural network was introduced into the pattern recognition of the intrusion signals in the optical fiber perimeter defense zone, and the different characteristics of the intrusion signals were extracted, so as to realize the accurate identification of different intrusion signals. Experimental results showed that this method was highly sensitive to intrusion events, could effectively reduce the false alarm rate of intrusion signals, and could improve the accuracy and efficiency of intrusion signal recognition.

Deep Spiking Neural Network for Video-Based Disguise Face Recognition Based on Dynamic Facial Movements

With the increasing popularity of social media and smart devices, the face as one of the key biometrics becomes vital for person identification. Among those face recognition algorithms, video-based face recognition methods could make use of both temporal and spatial information just as humans do to achieve better classification performance. However, they cannot identify individuals when certain key facial areas, such as eyes or nose, are disguised by heavy makeup or rubber/digital masks. To this end, we propose a novel deep spiking neural network architecture in this paper. It takes dynamic facial movements, the facial muscle changes induced by speaking or other activities, as the sole input. An event-driven continuous spike-timing-dependent plasticity learning rule with adaptive thresholding is applied to train the synaptic weights. The experiments on our proposed video-based disguise face database (MakeFace DB) demonstrate that the proposed learning method performs very well, i.e., it achieves from 95% to 100% correct classification rates under various realistic experimental scenarios.

Reversible uptake and release of sodium ions in layered SnS2-reduced graphene oxide composites for neuromorphic devices

With the advent of brain-inspired computing for complex data processing, emerging nonvolatile memories have been widely studied to develop neuromorphic devices for pattern recognition and deep learning. However, the devices still suffer from limitations such as nonlinearity and large write noise because they adopt a stochastic switching approach. Here, we suggest a biomimetic three-terminal electrochemical artificial synapse that is operated by a conductance change in response to intercalation of sodium (Na+) ions into a layered SnS2-reduced graphene oxide (RGO) composite channel. SnS2-RGO can reversibly uptake and release Na+ ions, so the conductance of the channel in artificial synapse can be controlled effectively and thereby it can emulate essential synaptic functions including short-term plasticity, spatiotemporal signal processing, and transition from short-term to long-term plasticity. The artificial synapse also shows linear and symmetric potentiation/depression with low cycle-to-cycle variation; these responses could improve the write linearity and reduce the write noise of devices. This study demonstrates the feasibility of next-generation neuromorphic memory using ion-based electrochemical devices that can mimic biological synapses with the migration of Na+ ions.

Memristive Spiking Neural Networks Trained with Unsupervised STDP

Neuromorphic computing systems are promising alternatives in the fields of pattern recognition, image processing, etc. especially when conventional von Neumann architectures face several bottlenecks. Memristors play vital roles in neuromorphic computing systems and are usually used as synaptic devices. Memristive spiking neural networks (MSNNs) are considered to be more efficient and biologically plausible than other systems due to their spike-based working mechanism. In contrast to previous SNNs with complex architectures, we propose a hardware-friendly architecture and an unsupervised spike-timing dependent plasticity (STDP) learning method for MSNNs in this paper. The architecture, which is friendly to hardware implementation, includes an input layer, a feature learning layer and a voting circuit. To reduce hardware complexity, some constraints are enforced: the proposed architecture has no lateral inhibition and is purely feedforward; it uses the voting circuit as a classifier and does not use additional classifiers; all neurons can generate at most one spike and do not need to consider firing rates and refractory periods; all neurons have the same fixed threshold voltage for classification. The presented unsupervised STDP learning method is time-dependent and uses no homeostatic mechanism. The MNIST dataset is used to demonstrate our proposed architecture and learning method. Simulation results show that our proposed architecture with the learning method achieves a classification accuracy of 94.6%, which outperforms other unsupervised SNNs that use time-based encoding schemes.