An Event-Driven Classifier for Spiking Neural Networks Fed with Synthetic or Dynamic Vision Sensor Data

Memristor-CMOS Hybrid Circuits Implementing Event-Driven Neural Networks for Dynamic Vision Sensor Camera

For processing streaming events from a Dynamic Vision Sensor camera, two types of neural networks can be considered. One are spiking neural networks, where simple spike-based computation is suitable for low-power consumption, but the discontinuity in spikes can make the training complicated in terms of hardware. The other one are digital Complementary Metal Oxide Semiconductor (CMOS)-based neural networks that can be trained directly using the normal backpropagation algorithm. However, the hardware and energy overhead can be significantly large, because all streaming events must be accumulated and converted into histogram data, which requires a large amount of memory such as SRAM. In this paper, to combine the spike-based operation with the normal backpropagation algorithm, memristor-CMOS hybrid circuits are proposed for implementing event-driven neural networks in hardware. The proposed hybrid circuits are composed of input neurons, synaptic crossbars, hidden/output neurons, and a neural network's controller. Firstly, the input neurons perform preprocessing for the DVS camera's events. The events are converted to histogram data using very simple memristor-based latches in the input neurons. After preprocessing the events, the converted histogram data are delivered to an ANN implemented using synaptic memristor crossbars. The memristor crossbars can perform low-power Multiply-Accumulate (MAC) calculations according to the memristor's current-voltage relationship. The hidden and output neurons can convert the crossbar's column currents to the output voltages according to the Rectified Linear Unit (ReLU) activation function. The neural network's controller adjusts the MAC calculation frequency according to the workload of the event computation. Moreover, the controller can disable the MAC calculation clock automatically to minimize unnecessary power consumption. The proposed hybrid circuits have been verified by circuit simulation for several event-based datasets such as POKER-DVS and MNIST-DVS. The circuit simulation results indicate that the neural network's performance proposed in this paper is degraded by as low as 0.5% while saving as much as 79% in power consumption for POKER-DVS. The recognition rate of the proposed scheme is lower by 0.75% compared to the conventional one, for the MNIST-DVS dataset. In spite of this little loss, the power consumption can be reduced by as much as 75% for the proposed scheme.

SHIP: a computational framework for simulating and validating novel technologies in hardware spiking neural networks

Investigations in the field of spiking neural networks (SNNs) encompass diverse, yet overlapping, scientific disciplines. Examples range from purely neuroscientific investigations, researches on computational aspects of neuroscience, or applicative-oriented studies aiming to improve SNNs performance or to develop artificial hardware counterparts. However, the simulation of SNNs is a complex task that can not be adequately addressed with a single platform applicable to all scenarios. The optimization of a simulation environment to meet specific metrics often entails compromises in other aspects. This computational challenge has led to an apparent dichotomy of approaches, with model-driven algorithms dedicated to the detailed simulation of biological networks, and data-driven algorithms designed for efficient processing of large input datasets. Nevertheless, material scientists, device physicists, and neuromorphic engineers who develop new technologies for spiking neuromorphic hardware solutions would find benefit in a simulation environment that borrows aspects from both approaches, thus facilitating modeling, analysis, and training of prospective SNN systems. This manuscript explores the numerical challenges deriving from the simulation of spiking neural networks, and introduces SHIP, Spiking (neural network) Hardware In PyTorch, a numerical tool that supports the investigation and/or validation of materials, devices, small circuit blocks within SNN architectures. SHIP facilitates the algorithmic definition of the models for the components of a network, the monitoring of states and output of the modeled systems, and the training of the synaptic weights of the network, by way of user-defined unsupervised learning rules or supervised training techniques derived from conventional machine learning. SHIP offers a valuable tool for researchers and developers in the field of hardware-based spiking neural networks, enabling efficient simulation and validation of novel technologies.

Indium-Gallium-Zinc Oxide-Based Synaptic Charge Trap Flash for Spiking Neural Network-Restricted Boltzmann Machine

Recently, neuromorphic computing has been proposed to overcome the drawbacks of the current von Neumann computing architecture. Especially, spiking neural network (SNN) has received significant attention due to its ability to mimic the spike-driven behavior of biological neurons and synapses, potentially leading to low-power consumption and other advantages. In this work, we designed the indium-gallium-zinc oxide (IGZO) channel charge-trap flash (CTF) synaptic device based on a HfO2/Al2O3/Si3N4/Al2O3 layer. Our IGZO-based CTF device exhibits synaptic functions with 128 levels of synaptic weight states and spike-timing-dependent plasticity. The SNN-restricted Boltzmann machine was used to simulate the fabricated CTF device to evaluate the efficiency for the SNN system, achieving the high pattern-recognition accuracy of 83.9%. We believe that our results show the suitability of the fabricated IGZO CTF device as a synaptic device for neuromorphic computing.

LIAF-Net: Leaky Integrate and Analog Fire Network for Lightweight and Efficient Spatiotemporal Information Processing

Spiking neural networks (SNNs) based on the leaky integrate and fire (LIF) model have been applied to energy-efficient temporal and spatiotemporal processing tasks. Due to the bioplausible neuronal dynamics and simplicity, LIF-SNN benefits from event-driven processing, however, usually face the embarrassment of reduced performance. This may because, in LIF-SNN, the neurons transmit information via spikes. To address this issue, in this work, we propose a leaky integrate and analog fire (LIAF) neuron model so that analog values can be transmitted among neurons, and a deep network termed LIAF-Net is built on it for efficient spatiotemporal processing. In the temporal domain, LIAF follows the traditional LIF dynamics to maintain its temporal processing capability. In the spatial domain, LIAF is able to integrate spatial information through convolutional integration or fully connected integration. As a spatiotemporal layer, LIAF can also be used with traditional artificial neural network (ANN) layers jointly. In addition, the built network can be trained with backpropagation through time (BPTT) directly, which avoids the performance loss caused by ANN to SNN conversion. Experiment results indicate that LIAF-Net achieves comparable performance to the gated recurrent unit (GRU) and long short-term memory (LSTM) on bAbI question answering (QA) tasks and achieves state-of-the-art performance on spatiotemporal dynamic vision sensor (DVS) data sets, including MNIST-DVS, CIFAR10-DVS, and DVS128 Gesture, with much less number of synaptic weights and computational overhead compared with traditional networks built by LSTM, GRU, convolutional LSTM (ConvLSTM), or 3-D convolution (Conv3D). Compared with traditional LIF-SNN, LIAF-Net also shows dramatic accuracy gain on all these experiments. In conclusion, LIAF-Net provides a framework combining the advantages of both ANNs and SNNs for lightweight and efficient spatiotemporal information processing.

Event-Driven Intrinsic Plasticity for Spiking Convolutional Neural Networks

The biologically discovered intrinsic plasticity (IP) learning rule, which changes the intrinsic excitability of an individual neuron by adaptively turning the firing threshold, has been shown to be crucial for efficient information processing. However, this learning rule needs extra time for updating operations at each step, causing extra energy consumption and reducing the computational efficiency. The event-driven or spike-based coding strategy of spiking neural networks (SNNs), i.e., neurons will only be active if driven by continuous spiking trains, employs all-or-none pulses (spikes) to transmit information, contributing to sparseness in neuron activations. In this article, we propose two event-driven IP learning rules, namely, input-driven and self-driven IP, based on basic IP learning. Input-driven means that IP updating occurs only when the neuron receives spiking inputs from its presynaptic neurons, whereas self-driven means that IP updating only occurs when the neuron generates a spike. A spiking convolutional neural network (SCNN) is developed based on the ANN2SNN conversion method, i.e., converting a well-trained rate-based artificial neural network to an SNN via directly mapping the connection weights. By comparing the computational performance of SCNNs with different IP rules on the recognition of MNIST, FashionMNIST, Cifar10, and SVHN datasets, we demonstrate that the two event-based IP rules can remarkably reduce IP updating operations, contributing to sparse computations and accelerating the recognition process. This work may give insights into the modeling of brain-inspired SNNs for low-power applications.

Advancements in Algorithms and Neuromorphic Hardware for Spiking Neural Networks

Artificial neural networks (ANNs) have experienced a rapid advancement for their success in various application domains, including autonomous driving and drone vision. Researchers have been improving the performance efficiency and computational requirement of ANNs inspired by the mechanisms of the biological brain. Spiking neural networks (SNNs) provide a power-efficient and brain-inspired computing paradigm for machine learning applications. However, evaluating large-scale SNNs on classical von Neumann architectures (central processing units/graphics processing units) demands a high amount of power and time. Therefore, hardware designers have developed neuromorphic platforms to execute SNNs in and approach that combines fast processing and low power consumption. Recently, field-programmable gate arrays (FPGAs) have been considered promising candidates for implementing neuromorphic solutions due to their varied advantages, such as higher flexibility, shorter design, and excellent stability. This review aims to describe recent advances in SNNs and the neuromorphic hardware platforms (digital, analog, hybrid, and FPGA based) suitable for their implementation. We present that biological background of SNN learning, such as neuron models and information encoding techniques, followed by a categorization of SNN training. In addition, we describe state-of-the-art SNN simulators. Furthermore, we review and present FPGA-based hardware implementation of SNNs. Finally, we discuss some future directions for research in this field.

Event-Based Vision: A Survey

Event cameras are bio-inspired sensors that differ from conventional frame cameras: Instead of capturing images at a fixed rate, they asynchronously measure per-pixel brightness changes, and output a stream of events that encode the time, location and sign of the brightness changes. Event cameras offer attractive properties compared to traditional cameras: high temporal resolution (in the order of μs), very high dynamic range (140 dB versus 60 dB), low power consumption, and high pixel bandwidth (on the order of kHz) resulting in reduced motion blur. Hence, event cameras have a large potential for robotics and computer vision in challenging scenarios for traditional cameras, such as low-latency, high speed, and high dynamic range. However, novel methods are required to process the unconventional output of these sensors in order to unlock their potential. This paper provides a comprehensive overview of the emerging field of event-based vision, with a focus on the applications and the algorithms developed to unlock the outstanding properties of event cameras. We present event cameras from their working principle, the actual sensors that are available and the tasks that they have been used for, from low-level vision (feature detection and tracking, optic flow, etc.) to high-level vision (reconstruction, segmentation, recognition). We also discuss the techniques developed to process events, including learning-based techniques, as well as specialized processors for these novel sensors, such as spiking neural networks. Additionally, we highlight the challenges that remain to be tackled and the opportunities that lie ahead in the search for a more efficient, bio-inspired way for machines to perceive and interact with the world.

Spiking Neural Network-Based Near-Sensor Computing for Damage Detection in Structural Health Monitoring

In this work, we present an innovative approach for damage detection of infrastructures on-edge devices, exploiting a brain-inspired algorithm. The proposed solution exploits recurrent spiking neural networks (LSNNs), which are emerging for their theoretical energy efficiency and compactness, to recognise damage conditions by processing data from low-cost accelerometers (MEMS) directly on the sensor node. We focus on designing an efficient coding of MEMS data to optimise SNN execution on a low-power microcontroller. We characterised and profiled LSNN performance and energy consumption on a hardware prototype sensor node equipped with an STM32 embedded microcontroller and a digital MEMS accelerometer. We used a hardware-in-the-loop environment with virtual sensors generating data on an SPI interface connected to the physical microcontroller to evaluate the system with a data stream from a real viaduct. We exploited this environment also to study the impact of different on-sensor encoding techniques, mimicking a bio-inspired sensor able to generate events instead of accelerations. Obtained results show that the proposed optimised embedded LSNN (eLSNN), when using a spike-based input encoding technique, achieves 54% lower execution time with respect to a naive LSNN algorithm implementation present in the state-of-the-art. The optimised eLSNN requires around 47 kCycles, which is comparable with the data transfer cost from the SPI interface. However, the spike-based encoding technique requires considerably larger input vectors to get the same classification accuracy, resulting in a longer pre-processing and sensor access time. Overall the event-based encoding techniques leads to a longer execution time (1.49×) but similar energy consumption. Moving this coding on the sensor can remove this limitation leading to an overall more energy-efficient monitoring system.

A Survey of Encoding Techniques for Signal Processing in Spiking Neural Networks

Biologically inspired spiking neural networks are increasingly popular in the field of artificial intelligence due to their ability to solve complex problems while being power efficient. They do so by leveraging the timing of discrete spikes as main information carrier. Though, industrial applications are still lacking, partially because the question of how to encode incoming data into discrete spike events cannot be uniformly answered. In this paper, we summarise the signal encoding schemes presented in the literature and propose a uniform nomenclature to prevent the vague usage of ambiguous definitions. Therefore we survey both, the theoretical foundations as well as applications of the encoding schemes. This work provides a foundation in spiking signal encoding and gives an overview over different application-oriented implementations which utilise the schemes.

Event-Based Sensing and Signal Processing in the Visual, Auditory, and Olfactory Domain: A Review

The nervous systems converts the physical quantities sensed by its primary receptors into trains of events that are then processed in the brain. The unmatched efficiency in information processing has long inspired engineers to seek brain-like approaches to sensing and signal processing. The key principle pursued in neuromorphic sensing is to shed the traditional approach of periodic sampling in favor of an event-driven scheme that mimicks sampling as it occurs in the nervous system, where events are preferably emitted upon the change of the sensed stimulus. In this paper we highlight the advantages and challenges of event-based sensing and signal processing in the visual, auditory and olfactory domains. We also provide a survey of the literature covering neuromorphic sensing and signal processing in all three modalities. Our aim is to facilitate research in event-based sensing and signal processing by providing a comprehensive overview of the research performed previously as well as highlighting conceptual advantages, current progress and future challenges in the field.

A Heterogeneous Spiking Neural Network for Unsupervised Learning of Spatiotemporal Patterns

This paper introduces a heterogeneous spiking neural network (H-SNN) as a novel, feedforward SNN structure capable of learning complex spatiotemporal patterns with spike-timing-dependent plasticity (STDP) based unsupervised training. Within H-SNN, hierarchical spatial and temporal patterns are constructed with convolution connections and memory pathways containing spiking neurons with different dynamics. We demonstrate analytically the formation of long and short term memory in H-SNN and distinct response functions of memory pathways. In simulation, the network is tested on visual input of moving objects to simultaneously predict for object class and motion dynamics. Results show that H-SNN achieves prediction accuracy on similar or higher level than supervised deep neural networks (DNN). Compared to SNN trained with back-propagation, H-SNN effectively utilizes STDP to learn spatiotemporal patterns that have better generalizability to unknown motion and/or object classes encountered during inference. In addition, the improved performance is achieved with 6x fewer parameters than complex DNNs, showing H-SNN as an efficient approach for applications with constrained computation resources.

Perception Understanding Action: Adding Understanding to the Perception Action Cycle With Spiking Segmentation

Traditionally the Perception Action cycle is the first stage of building an autonomous robotic system and a practical way to implement a low latency reactive system within a low Size, Weight and Power (SWaP) package. However, within complex scenarios, this method can lack contextual understanding about the scene, such as object recognition-based tracking or system attention. Object detection, identification and tracking along with semantic segmentation and attention are all modern computer vision tasks in which Convolutional Neural Networks (CNN) have shown significant success, although such networks often have a large computational overhead and power requirements, which are not ideal in smaller robotics tasks. Furthermore, cloud computing and massively parallel processing like in Graphic Processing Units (GPUs) are outside the specification of many tasks due to their respective latency and SWaP constraints. In response to this, Spiking Convolutional Neural Networks (SCNNs) look to provide the feature extraction benefits of CNNs, while maintaining low latency and power overhead thanks to their asynchronous spiking event-based processing. A novel Neuromorphic Perception Understanding Action (PUA) system is presented, that aims to combine the feature extraction benefits of CNNs with low latency processing of SCNNs. The PUA utilizes a Neuromorphic Vision Sensor for Perception that facilitates asynchronous processing within a Spiking fully Convolutional Neural Network (SpikeCNN) to provide semantic segmentation and Understanding of the scene. The output is fed to a spiking control system providing Actions. With this approach, the aim is to bring features of deep learning into the lower levels of autonomous robotics, while maintaining a biologically plausible STDP rule throughout the learned encoding part of the network. The network will be shown to provide a more robust and predictable management of spiking activity with an improved thresholding response. The reported experiments show that this system can deliver robust results of over 96 and 81% for accuracy and Intersection over Union, ensuring such a system can be successfully used within object recognition, classification and tracking problem. This demonstrates that the attention of the system can be tracked accurately, while the asynchronous processing means the controller can give precise track updates with minimal latency.

Analogue pattern recognition with stochastic switching binary CMOS-integrated memristive devices

Biological neural networks outperform current computer technology in terms of power consumption and computing speed while performing associative tasks, such as pattern recognition. The analogue and massive parallel in-memory computing in biology differs strongly from conventional transistor electronics that rely on the von Neumann architecture. Therefore, novel bio-inspired computing architectures have been attracting a lot of attention in the field of neuromorphic computing. Here, memristive devices, which serve as non-volatile resistive memory, are employed to emulate the plastic behaviour of biological synapses. In particular, CMOS integrated resistive random access memory (RRAM) devices are promising candidates to extend conventional CMOS technology to neuromorphic systems. However, dealing with the inherent stochasticity of resistive switching can be challenging for network performance. In this work, the probabilistic switching is exploited to emulate stochastic plasticity with fully CMOS integrated binary RRAM devices. Two different RRAM technologies with different device variabilities are investigated in detail, and their potential applications in stochastic artificial neural networks (StochANNs) capable of solving MNIST pattern recognition tasks is examined. A mixed-signal implementation with hardware synapses and software neurons combined with numerical simulations shows that the proposed concept of stochastic computing is able to process analogue data with binary memory cells.

Bio-Inspired Evolutionary Model of Spiking Neural Networks in Ionic Liquid Space

One of the biggest struggles while working with artificial neural networks is being able to come up with models which closely match biological observations. Biological neural networks seem to capable of creating and pruning dendritic spines, leading to synapses being changed, which results in higher learning capability. The latter forms the basis of the present study in which a new ionic model for reservoir-like networks, consisting of spiking neurons, is introduced. High plasticity of this model makes learning possible with a fewer number of neurons. In order to study the effect of the applied stimulus in an ionic liquid space through time, a diffusion operator is used which somehow compensates for the separation between spatial and temporal coding in spiking neural networks and therefore, makes the mentioned model suitable for spatiotemporal patterns. Inspired by partial structural changes in the human brain over the years, the proposed model evolves during the learning process. The effect of topological evolution on the proposed model's performance for some classification problems is studied in this paper. Several datasets have been used to evaluate the performance of the proposed model compared to the original LSM. Classification results via separation and accuracy values have shown that the proposed ionic liquid outperforms the original LSM.

Towards spike-based machine intelligence with neuromorphic computing

Guided by brain-like 'spiking' computational frameworks, neuromorphic computing-brain-inspired computing for machine intelligence-promises to realize artificial intelligence while reducing the energy requirements of computing platforms. This interdisciplinary field began with the implementation of silicon circuits for biological neural routines, but has evolved to encompass the hardware implementation of algorithms with spike-based encoding and event-driven representations. Here we provide an overview of the developments in neuromorphic computing for both algorithms and hardware and highlight the fundamentals of learning and hardware frameworks. We discuss the main challenges and the future prospects of neuromorphic computing, with emphasis on algorithm-hardware codesign.

ReStoCNet: Residual Stochastic Binary Convolutional Spiking Neural Network for Memory-Efficient Neuromorphic Computing

In this work, we propose ReStoCNet, a residual stochastic multilayer convolutional Spiking Neural Network (SNN) composed of binary kernels, to reduce the synaptic memory footprint and enhance the computational efficiency of SNNs for complex pattern recognition tasks. ReStoCNet consists of an input layer followed by stacked convolutional layers for hierarchical input feature extraction, pooling layers for dimensionality reduction, and fully-connected layer for inference. In addition, we introduce residual connections between the stacked convolutional layers to improve the hierarchical feature learning capability of deep SNNs. We propose Spike Timing Dependent Plasticity (STDP) based probabilistic learning algorithm, referred to as Hybrid-STDP (HB-STDP), incorporating Hebbian and anti-Hebbian learning mechanisms, to train the binary kernels forming ReStoCNet in a layer-wise unsupervised manner. We demonstrate the efficacy of ReStoCNet and the presented HB-STDP based unsupervised training methodology on the MNIST and CIFAR-10 datasets. We show that residual connections enable the deeper convolutional layers to self-learn useful high-level input features and mitigate the accuracy loss observed in deep SNNs devoid of residual connections. The proposed ReStoCNet offers >20 × kernel memory compression compared to full-precision (32-bit) SNN while yielding high enough classification accuracy on the chosen pattern recognition tasks.

Deep learning in spiking neural networks

In recent years, deep learning has revolutionized the field of machine learning, for computer vision in particular. In this approach, a deep (multilayer) artificial neural network (ANN) is trained, most often in a supervised manner using backpropagation. Vast amounts of labeled training examples are required, but the resulting classification accuracy is truly impressive, sometimes outperforming humans. Neurons in an ANN are characterized by a single, static, continuous-valued activation. Yet biological neurons use discrete spikes to compute and transmit information, and the spike times, in addition to the spike rates, matter. Spiking neural networks (SNNs) are thus more biologically realistic than ANNs, and are arguably the only viable option if one wants to understand how the brain computes at the neuronal description level. The spikes of biological neurons are sparse in time and space, and event-driven. Combined with bio-plausible local learning rules, this makes it easier to build low-power, neuromorphic hardware for SNNs. However, training deep SNNs remains a challenge. Spiking neurons' transfer function is usually non-differentiable, which prevents using backpropagation. Here we review recent supervised and unsupervised methods to train deep SNNs, and compare them in terms of accuracy and computational cost. The emerging picture is that SNNs still lag behind ANNs in terms of accuracy, but the gap is decreasing, and can even vanish on some tasks, while SNNs typically require many fewer operations and are the better candidates to process spatio-temporal data.

Deep Learning With Spiking Neurons: Opportunities and Challenges

Spiking neural networks (SNNs) are inspired by information processing in biology, where sparse and asynchronous binary signals are communicated and processed in a massively parallel fashion. SNNs on neuromorphic hardware exhibit favorable properties such as low power consumption, fast inference, and event-driven information processing. This makes them interesting candidates for the efficient implementation of deep neural networks, the method of choice for many machine learning tasks. In this review, we address the opportunities that deep spiking networks offer and investigate in detail the challenges associated with training SNNs in a way that makes them competitive with conventional deep learning, but simultaneously allows for efficient mapping to hardware. A wide range of training methods for SNNs is presented, ranging from the conversion of conventional deep networks into SNNs, constrained training before conversion, spiking variants of backpropagation, and biologically motivated variants of STDP. The goal of our review is to define a categorization of SNN training methods, and summarize their advantages and drawbacks. We further discuss relationships between SNNs and binary networks, which are becoming popular for efficient digital hardware implementation. Neuromorphic hardware platforms have great potential to enable deep spiking networks in real-world applications. We compare the suitability of various neuromorphic systems that have been developed over the past years, and investigate potential use cases. Neuromorphic approaches and conventional machine learning should not be considered simply two solutions to the same classes of problems, instead it is possible to identify and exploit their task-specific advantages. Deep SNNs offer great opportunities to work with new types of event-based sensors, exploit temporal codes and local on-chip learning, and we have so far just scratched the surface of realizing these advantages in practical applications.

On Practical Issues for Stochastic STDP Hardware With 1-bit Synaptic Weights

In computational neuroscience, synaptic plasticity learning rules are typically studied using the full 64-bit floating point precision computers provide. However, for dedicated hardware implementations, the precision used not only penalizes directly the required memory resources, but also the computing, communication, and energy resources. When it comes to hardware engineering, a key question is always to find the minimum number of necessary bits to keep the neurocomputational system working satisfactorily. Here we present some techniques and results obtained when limiting synaptic weights to 1-bit precision, applied to a Spike-Timing-Dependent-Plasticity (STDP) learning rule in Spiking Neural Networks (SNN). We first illustrate the 1-bit synapses STDP operation by replicating a classical biological experiment on visual orientation tuning, using a simple four neuron setup. After this, we apply 1-bit STDP learning to the hidden feature extraction layer of a 2-layer system, where for the second (and output) layer we use already reported SNN classifiers. The systems are tested on two spiking datasets: a Dynamic Vision Sensor (DVS) recorded poker card symbols dataset and a Poisson-distributed spike representation MNIST dataset version. Tests are performed using the in-house MegaSim event-driven behavioral simulator and by implementing the systems on FPGA (Field Programmable Gate Array) hardware.

Training Deep Spiking Convolutional Neural Networks With STDP-Based Unsupervised Pre-training Followed by Supervised Fine-Tuning

Spiking Neural Networks (SNNs) are fast becoming a promising candidate for brain-inspired neuromorphic computing because of their inherent power efficiency and impressive inference accuracy across several cognitive tasks such as image classification and speech recognition. The recent efforts in SNNs have been focused on implementing deeper networks with multiple hidden layers to incorporate exponentially more difficult functional representations. In this paper, we propose a pre-training scheme using biologically plausible unsupervised learning, namely Spike-Timing-Dependent-Plasticity (STDP), in order to better initialize the parameters in multi-layer systems prior to supervised optimization. The multi-layer SNN is comprised of alternating convolutional and pooling layers followed by fully-connected layers, which are populated with leaky integrate-and-fire spiking neurons. We train the deep SNNs in two phases wherein, first, convolutional kernels are pre-trained in a layer-wise manner with unsupervised learning followed by fine-tuning the synaptic weights with spike-based supervised gradient descent backpropagation. Our experiments on digit recognition demonstrate that the STDP-based pre-training with gradient-based optimization provides improved robustness, faster (~2.5 ×) training time and better generalization compared with purely gradient-based training without pre-training.

Active Perception With Dynamic Vision Sensors. Minimum Saccades With Optimum Recognition

Vision processing with dynamic vision sensors (DVSs) is becoming increasingly popular. This type of a bio-inspired vision sensor does not record static images. The DVS pixel activity relies on the changes in light intensity. In this paper, we introduce a platform for the object recognition with a DVS in which the sensor is installed on a moving pan-tilt unit in a closed loop with a recognition neural network. This neural network is trained to recognize objects observed by a DVS, while the pan-tilt unit is moved to emulate micro-saccades. We show that performing more saccades in different directions can result in having more information about the object, and therefore, more accurate object recognition is possible. However, in high-performance and low-latency platforms, performing additional saccades adds latency and power consumption. Here, we show that the number of saccades can be reduced while keeping the same recognition accuracy by performing intelligent saccadic movements, in a closed action-perception smart loop. We propose an algorithm for smart saccadic movement decisions that can reduce the number of necessary saccades to half, on average, for a predefined accuracy on the N-MNIST dataset. Additionally, we show that by replacing this control algorithm with an artificial neural network that learns to control the saccades, we can also reduce to half the average number of saccades needed for the N-MNIST recognition.

Event-Based, Timescale Invariant Unsupervised Online Deep Learning With STDP

Learning of hierarchical features with spiking neurons has mostly been investigated in the database framework of standard deep learning systems. However, the properties of neuromorphic systems could be particularly interesting for learning from continuous sensor data in real-world settings. In this work, we introduce a deep spiking convolutional neural network of integrate-and-fire (IF) neurons which performs unsupervised online deep learning with spike-timing dependent plasticity (STDP) from a stream of asynchronous and continuous event-based data. In contrast to previous approaches to unsupervised deep learning with spikes, where layers were trained successively, we introduce a mechanism to train all layers of the network simultaneously. This allows approximate online inference already during the learning process and makes our architecture suitable for online learning and inference. We show that it is possible to train the network without providing implicit information about the database, such as the number of classes and the duration of stimuli presentation. By designing an STDP learning rule which depends only on relative spike timings, we make our network fully event-driven and able to operate without defining an absolute timescale of its dynamics. Our architecture requires only a small number of generic mechanisms and therefore enforces few constraints on a possible neuromorphic hardware implementation. These characteristics make our network one of the few neuromorphic architecture which could directly learn features and perform inference from an event-based vision sensor.

A Retinotopic Spiking Neural Network System for Accurate Recognition of Moving Objects Using NeuCube and Dynamic Vision Sensors

This paper introduces a new system for dynamic visual recognition that combines bio-inspired hardware with a brain-like spiking neural network. The system is designed to take data from a dynamic vision sensor (DVS) that simulates the functioning of the human retina by producing an address event output (spike trains) based on the movement of objects. The system then convolutes the spike trains and feeds them into a brain-like spiking neural network, called NeuCube, which is organized in a three-dimensional manner, representing the organization of the primary visual cortex. Spatio-temporal patterns of the data are learned during a deep unsupervised learning stage, using spike-timing-dependent plasticity. In a second stage, supervised learning is performed to train the network for classification tasks. The convolution algorithm and the mapping into the network mimic the function of retinal ganglion cells and the retinotopic organization of the visual cortex. The NeuCube architecture can be used to visualize the deep connectivity inside the network before, during, and after training and thereby allows for a better understanding of the learning processes. The method was tested on the benchmark MNIST-DVS dataset and achieved a classification accuracy of 92.90%. The paper discusses advantages and limitations of the new method and concludes that it is worth exploring further on different datasets, aiming for advances in dynamic computer vision and multimodal systems that integrate visual, aural, tactile, and other kinds of information in a biologically plausible way.

Conversion of Continuous-Valued Deep Networks to Efficient Event-Driven Networks for Image Classification

Spiking neural networks (SNNs) can potentially offer an efficient way of doing inference because the neurons in the networks are sparsely activated and computations are event-driven. Previous work showed that simple continuous-valued deep Convolutional Neural Networks (CNNs) can be converted into accurate spiking equivalents. These networks did not include certain common operations such as max-pooling, softmax, batch-normalization and Inception-modules. This paper presents spiking equivalents of these operations therefore allowing conversion of nearly arbitrary CNN architectures. We show conversion of popular CNN architectures, including VGG-16 and Inception-v3, into SNNs that produce the best results reported to date on MNIST, CIFAR-10 and the challenging ImageNet dataset. SNNs can trade off classification error rate against the number of available operations whereas deep continuous-valued neural networks require a fixed number of operations to achieve their classification error rate. From the examples of LeNet for MNIST and BinaryNet for CIFAR-10, we show that with an increase in error rate of a few percentage points, the SNNs can achieve more than 2x reductions in operations compared to the original CNNs. This highlights the potential of SNNs in particular when deployed on power-efficient neuromorphic spiking neuron chips, for use in embedded applications.