Search

Spiking neural network (SNN) is interesting due to its strong bio-plausibility and high energy efficiency. However, its performance is falling far behind conventional deep neural networks (DNNs). In this paper, considering a general class of single-spike temporal-coded integrate-and-fire neurons, we analyze the input-output expressions of both leaky and nonleaky neurons. We show that SNNs built with leaky neurons suffer from the overly-nonlinear and overly-complex input-output response, which is the major reason for their difficult training and low performance. This reason is more fundamental than the commonly believed problem of nondifferentiable spikes. To support this claim, we show that SNNs built with nonleaky neurons can have a less-complex and less-nonlinear input-output response. They can be easily trained and can have superior performance, which is demonstrated by experimenting with the SNNs over two popular network intrusion detection datasets, i.e., the NSL-KDD and the AWID datasets. Our experiment results show that the proposed SNNs outperform a comprehensive list of DNN models and classic machine learning models. This paper demonstrates that SNNs can be promising and competitive in contrast to common beliefs., Comment: Published in the 25th International Conference on Pattern Recognition (ICPR'2020), January 2021

Image sensors have been explored heavily in automotive applications for collision avoidance and varying levels of autonomy. It requires a degree of brightness, therefore, the use of an image sensor in nighttime operation or dark conditions can be problematic along with challenging weather such as fog. Radar sensors have been employed to help cover the various environmental challenges with visible spectrum cameras. Edge computing technology has the potential to address a number of issues such as real-time processing requirements, off-loading of processing from congested servers, and size, weight, power, and cost (SWaP-C) constraints. This paper proposes a novel Hybrid Object DEtection and Tracking (HODET) using mmWave Radar and Visual Sensors at the edge. The HODET is a computing application of low SWaP-C electronics performing object detection, tracking and identification algorithms with the simultaneous use of image and radar sensors. While the machine vision camera alone could estimate the distance of an object, the radar sensor will provide an accurate distance and vector of movement. This additional data accuracy can be leveraged to further discriminate a detected object to protect against spoofing attacks. A real-world smart community public safety monitoring scenario is selected to verify the effectiveness of HODET, which detects, tracks objects of interests and identify suspicious activities. The experimental results demonstrate the feasibility of the approach., Comment: 2020 SPIE Defense + Commercial Sensing

Accurate real-time object recognition from sensory data has long been a crucial and challenging task for autonomous driving. Even though deep neural networks (DNNs) have been successfully applied in this area, most existing methods still heavily rely on the pre-processing of the pulse signals derived from LiDAR sensors, and therefore introduce additional computational overhead and considerable latency. In this paper, we propose an approach to address the object recognition problem directly with raw temporal pulses utilizing the spiking neural network (SNN). Being evaluated on various datasets (including Sim LiDAR, KITTI and DVS-barrel) derived from LiDAR and dynamic vision sensor (DVS), our proposed method has shown comparable performance as the state-of-the-art methods, while achieving remarkable time efficiency. It highlights the SNN's great potentials in autonomous driving and related applications. To the best of our knowledge, this is the first attempt to use SNN to directly perform object recognition on raw temporal pulses.

Real-time accurate detection of three-dimensional (3D) objects is a fundamental necessity for self-driving vehicles. Most existing computer vision approaches are based on convolutional neural networks (CNNs). Although the CNN-based approaches can achieve high detection accuracy, their high energy consumption is a severe drawback. To resolve this problem, novel energy efficient approaches should be explored. Spiking neural network (SNN) is a promising candidate because it has orders-of-magnitude lower energy consumption than CNN. Unfortunately, the studying of SNN has been limited in small networks only. The application of SNN for large 3D object detection networks has remain largely open. In this paper, we integrate spiking convolutional neural network (SCNN) with temporal coding into the YOLOv2 architecture for real-time object detection. To take the advantage of spiking signals, we develop a novel data preprocessing layer that translates 3D point-cloud data into spike time data. We propose an analog circuit to implement the non-leaky integrate and fire neuron used in our SCNN, from which the energy consumption of each spike is estimated. Moreover, we present a method to calculate the network sparsity and the energy consumption of the overall network. Extensive experiments have been conducted based on the KITTI dataset, which show that the proposed network can reach competitive detection accuracy as existing approaches, yet with much lower average energy consumption. If implemented in dedicated hardware, our network could have a mean sparsity of 56.24% and extremely low total energy consumption of 0.247mJ only. Implemented in NVIDIA GTX 1080i GPU, we can achieve 35.7 fps frame rate, high enough for real-time object detection.

The $\theta_K$ has been re-evaluated via mass relations and latest experimental results, meanwhile, the $M_{K_{1B}}$ also be obtained. Based on the singlet-octet mixing model and quark-flavor mixing model, the $\theta_{h_1}$ has been recalculated with a modified formula, by inputting $M_{K_{1B}}$ instead of $\theta_K$. The values are calculated to be $|{\theta _K}| = {(42.6 \pm 2.2)^ \circ }$, $M_{K_{1B}} = 1333.9 \pm 4.6\ {\rm{MeV}}/c^{2}$ and $\theta _{h_1} = 29.5\pm 2.0^\circ$. Using these calculations as input, we predict that the branching fraction of $J/\psi \to \eta'(\eta)h_1(1170)$ is about one order higher than that of $J/\psi \to \eta'(\eta)h_1(1415)$, which can be measured in future experiments to test the validity of these two models., Comment: 6 pages, with no figure

Spiking neural network (SNN) is interesting both theoretically and practically because of its strong bio-inspiration nature and potentially outstanding energy efficiency. Unfortunately, its development has fallen far behind the conventional deep neural network (DNN), mainly because of difficult training and lack of widely accepted hardware experiment platforms. In this paper, we show that a deep temporal-coded SNN can be trained easily and directly over the benchmark datasets CIFAR10 and ImageNet, with testing accuracy within 1% of the DNN of equivalent size and architecture. Training becomes similar to DNN thanks to the closed-form solution to the spiking waveform dynamics. Considering that SNNs should be implemented in practical neuromorphic hardwares, we train the deep SNN with weights quantized to 8, 4, 2 bits and with weights perturbed by random noise to demonstrate its robustness in practical applications. In addition, we develop a phase-domain signal processing circuit schematic to implement our spiking neuron with 90% gain of energy efficiency over existing work. This paper demonstrates that the temporal-coded deep SNN is feasible for applications with high performance and high energy efficient.