Feedforward Spiking Neural Network Research Articles

SpikeTemp: An Enhanced Rank-Order-Based Learning Approach for Spiking Neural Networks With Adaptive Structure.

This paper presents an enhanced rank-order-based learning algorithm, called SpikeTemp, for spiking neural networks (SNNs) with a dynamically adaptive structure. The trained feed-forward SNN consists of two layers of spiking neurons: 1) an encoding layer which temporally encodes real-valued features into spatio-temporal spike patterns and 2) an output layer of dynamically grown neurons which perform spatio-temporal classification. Both Gaussian receptive fields and square cosine population encoding schemes are employed to encode real-valued features into spatio-temporal spike patterns. Unlike the rank-order-based learning approach, SpikeTemp uses the precise times of the incoming spikes for adjusting the synaptic weights such that early spikes result in a large weight change and late spikes lead to a smaller weight change. This removes the need to rank all the incoming spikes and, thus, reduces the computational cost of SpikeTemp. The proposed SpikeTemp algorithm is demonstrated on several benchmark data sets and on an image recognition task. The results show that SpikeTemp can achieve better classification performance and is much faster than the existing rank-order-based learning approach. In addition, the number of output neurons is much smaller when the square cosine encoding scheme is employed. Furthermore, SpikeTemp is benchmarked against a selection of existing machine learning algorithms, and the results demonstrate the ability of SpikeTemp to classify different data sets after just one presentation of the training samples with comparable classification performance.

Bio-inspired unsupervised learning of visual features leads to robust invariant object recognition

Retinal image of surrounding objects varies tremendously due to the changes in position, size, pose, illumination condition, background context, occlusion, noise, and non-rigid deformations. But despite these huge variations, our visual system is able to invariantly recognize any object in just a fraction of a second. To date, various computational models have been proposed to mimic the hierarchical processing of the ventral visual pathway, with limited success. Here, we show that the association of both biologically inspired network architecture and learning rule significantly improves the models׳ performance when facing challenging invariant object recognition problems. Our model is an asynchronous feedforward spiking neural network. When the network is presented with natural images, the neurons in the entry layers detect edges, and the most activated ones fire first, while neurons in higher layers are equipped with spike timing-dependent plasticity. These neurons progressively become selective to intermediate complexity visual features appropriate for object categorization. The model is evaluated on 3D-Object and ETH-80 datasets which are two benchmarks for invariant object recognition, and is shown to outperform state-of-the-art models, including DeepConvNet and HMAX. This demonstrates its ability to accurately recognize different instances of multiple object classes even under various appearance conditions (different views, scales, tilts, and backgrounds). Several statistical analysis techniques are used to show that our model extracts class specific and highly informative features.

A Minimal Spiking Neural Network to Rapidly Train and Classify Handwritten Digits in Binary and 10-Digit Tasks

This paper reports the results of experiments to develop a minimal neural network for pattern classification. The network uses biologically plausible neural and learning mechanisms and is applied to a subset of the MNIST dataset of handwritten digits. The research goal is to assess the classification power of a very simple biologically motivated mechanism. The network architecture is primarily a feedforward spiking neural network (SNN) composed of Izhikevich regular spiking (RS) neurons and conductance-based synapses. The weights are trained with the spike timing-dependent plasticity (STDP) learning rule. The proposed SNN architecture contains three neuron layers which are connected by both static and adaptive synapses. Visual input signals are processed by the first layer to generate input spike trains. The second and third layers contribute to spike train segmentation and STDP learning, respectively. The network is evaluated by classification accuracy on the handwritten digit images from the MNIST dataset. The simulation results show that although the proposed SNN is trained quickly without error-feedbacks in a few number of iterations, it results in desirable performance (97.6%) in the binary classification (0 and 1). In addition, the proposed SNN gives acceptable recognition accuracy in 10-digit (0-9) classification in comparison with statistical methods such as support vector machine (SVM) and multi-perceptron neural network.

An online supervised learning method for spiking neural networks with adaptive structure

A novel online learning algorithm for Spiking Neural Networks (SNNs) with dynamically adaptive structure is presented. The main contribution of this work lies in the fact that the proposed adaptive SNN is able to classify spike-based spatio-temporal inputs after just one presentation of the training set, i.e. in one pass only, and does not require the entire training set to be available at once. Both the structure and weights of the SNN are learned dynamically through a combination of unsupervised and supervised learning paradigms. The proposed feed-forward SNN consists of three layers of spiking neurons: an input layer which temporally encodes real valued features into spike-based spatio-temporal patterns, a hidden layer of dynamically grown and pruned neurons which perform spatio-temporal clustering, and an output layer for classification. An unsupervised spiking-based clustering algorithm is implemented by the hidden layer whose spiking neurons are trained to compute a temporal Radial Basis Function (RBF) where incoming inputs will selectively activate hidden neurons based on how close the inputs are to the preferred inputs of the hidden neurons. The centre of each hidden RBF spiking neuron is represented by its time to first spike. In addition, a growing and pruning strategy is proposed to adjust the structure of the hidden layer ‘on-the-fly’ as inputs are presented to the SNN. Both the weights and the centres of the hidden RBF neurons are learned in an unsupervised way and classification at the output layer is achieved through supervised learning where the learning windows proposed for STDP and anti-STDP are used to adjust the weights of the output neurons afferent connections. Competition at both the hidden and the output layers is achieved through the use of lateral inhibitory connections between the neurons of each layer. The proposed online learning algorithm is validated on several benchmark datasets. The evaluation results demonstrate that SNNs trained with the proposed approach require only one pass through the training set in order to classify the inputs with comparable accuracies to existing SNN-based approaches as well as traditional representative classifiers.

Modular Neural Tile Architecture for Compact Embedded Hardware Spiking Neural Network

Biologically-inspired packet switched network on chip (NoC) based hardware spiking neural network (SNN) architectures have been proposed as an embedded computing platform for classification, estimation and control applications. Storage of large synaptic connectivity (SNN topology) information in SNNs require large distributed on-chip memory, which poses serious challenges for compact hardware implementation of such architectures. Based on the structured neural organisation observed in human brain, a modular neural networks (MNN) design strategy partitions complex application tasks into smaller subtasks executing on distinct neural network modules, and integrates intermediate outputs in higher level functions. This paper proposes a hardware modular neural tile (MNT) architecture that reduces the SNN topology memory requirement of NoC-based hardware SNNs by using a combination of fixed and configurable synaptic connections. The proposed MNT contains a 16:16 fully-connected feed-forward SNN structure and integrates in a mesh topology NoC communication infrastructure. The SNN topology memory requirement is 50 % of the monolithic NoC-based hardware SNN implementation. The paper also presents a lookup table based SNN topology memory allocation technique, which further increases the memory utilisation efficiency. Overall the area requirement of the architecture is reduced by an average of 66 % for practical SNN application topologies. The paper presents micro-architecture details of the proposed MNT and digital neuron circuit. The proposed architecture has been validated on a Xilinx Virtex-6 FPGA and synthesised using 65 nm low-power CMOS technology. The evolvable capability of the proposed MNT and its suitability for executing subtasks within a MNN execution architecture is demonstrated by successfully evolving benchmark SNN application tasks representing classification and non-linear control functions. The paper addresses hardware modular SNN design and implementation challenges and contributes to the development of a compact hardware modular SNN architecture suitable for embedded applications

Fast Learning in Spiking Neural Networks by Learning Rate Adaptation

For accelerating the supervised learning by the SpikeProp algorithm with the temporal coding paradigm in spiking neural networks (SNNs), three learning rate adaptation methods (heuristic rule, delta-delta rule, and delta-bar-delta rule), which are used to speed up training in artificial neural networks, are used to develop the training algorithms for feedforward SNN. The performance of these algorithms is investigated by four experiments: classical XOR (exclusive or) problem, Iris dataset, fault diagnosis in the Tennessee Eastman process, and Poisson trains of discrete spikes. The results demonstrate that all the three learning rate adaptation methods are able to speed up convergence of SNN compared with the original SpikeProp algorithm. Furthermore, if the adaptive learning rate is used in combination with the momentum term, the two modifications will balance each other in a beneficial way to accomplish rapid and steady convergence. In the three learning rate adaptation methods, delta-bar-delta rule performs the best. The delta-bar-delta method with momentum has the fastest convergence rate, the greatest stability of training process, and the maximum accuracy of network learning. The proposed algorithms in this paper are simple and efficient, and consequently valuable for practical applications of SNN.

Physical aspects of low power synapses based on phase change memory devices

In this work, we demonstrate how phase change memory (PCM) devices can be used to emulate biologically inspired synaptic functions in particular, potentiation and depression, important for implementing neuromorphic hardware. PCM devices with different chalcogenide materials are fabricated and characterized. The asymmetry between the potentiation and depression behaviors of the PCM is stressed. Detailed multi-physical simulations are performed to study the underlying physics of the synaptic behavior of PCM. A versatile behavioral model and a multi-level circuit-compatible model are developed for system and circuit-level neuromorphic simulations. We propose a unique low-power methodology named the 2-PCM Synapse, to use PCM devices as synapses in large scale neuromorphic systems. To show the strength of our proposed solution, we efficiently simulated fully connected feed-forward spiking neural network capable of complex visual pattern extraction from real world data.

Extraction of temporally correlated features from dynamic vision sensors with spike-timing-dependent plasticity

A biologically inspired approach to learning temporally correlated patterns from a spiking silicon retina is presented. Spikes are generated from the retina in response to relative changes in illumination at the pixel level and transmitted to a feed-forward spiking neural network. Neurons become sensitive to patterns of pixels with correlated activation times, in a fully unsupervised scheme. This is achieved using a special form of Spike-Timing-Dependent Plasticity which depresses synapses that did not recently contribute to the post-synaptic spike activation, regardless of their activation time. Competitive learning is implemented with lateral inhibition. When tested with real-life data, the system is able to extract complex and overlapping temporally correlated features such as car trajectories on a freeway, after only 10 min of traffic learning. Complete trajectories can be learned with a 98% detection rate using a second layer, still with unsupervised learning, and the system may be used as a car counter. The proposed neural network is extremely robust to noise and it can tolerate a high degree of synaptic and neuronal variability with little impact on performance. Such results show that a simple biologically inspired unsupervised learning scheme is capable of generating selectivity to complex meaningful events on the basis of relatively little sensory experience.

Unsupervised learning of visual features through spike timing dependent plasticity.

Spike timing dependent plasticity (STDP) is a learning rule that modifies synaptic strength as a function of the relative timing of pre- and postsynaptic spikes. When a neuron is repeatedly presented with similar inputs, STDP is known to have the effect of concentrating high synaptic weights on afferents that systematically fire early, while postsynaptic spike latencies decrease. Here we use this learning rule in an asynchronous feedforward spiking neural network that mimics the ventral visual pathway and shows that when the network is presented with natural images, selectivity to intermediate-complexity visual features emerges. Those features, which correspond to prototypical patterns that are both salient and consistently present in the images, are highly informative and enable robust object recognition, as demonstrated on various classification tasks. Taken together, these results show that temporal codes may be a key to understanding the phenomenal processing speed achieved by the visual system and that STDP can lead to fast and selective responses.

Sparse spike coding in an asynchronous feed-forward multi-layer neural network using matching pursuit

In order to account for the rapidity of visual processing, we explore visual coding strategies using a one-pass feed-forward spiking neural network. We based our model on the work of Van Rullen and Thorpe Neural Comput. 13 (6) (2001) 1255, which constructs a retinal representation using an orthogonal wavelet transform. This strategy provides a spike code, thanks to a rank order coding scheme which offers an alternative to the classical spike frequency coding scheme. We extended this model to efficient representations in arbitrary linear generative models by implementing lateral interactions on top of this feed-forward model. This method uses a matching pursuit scheme—recursively detecting in the image the best match with the elements of a dictionary and then subtracting it—and which may similarly define a visual spike code. In particular, this transform could be used with large and arbitrary dictionaries, so that we may define an over-complete representation which may define an efficient sparse spike coding scheme in arbitrary multi-layered architectures. We show here extensions of this method of computing with spike events, introducing an adaptive scheme leading to the emergence of V1-like receptive fields and then a model of bottom-up saliency pursuit.