ALBSNN: ultra-low latency adaptive local binary spiking neural network with accuracy loss estimator.

Spiking neural network (SNN) is a brain-inspired model with more spatio-temporal information processing capacity and computational energy efficiency. However, with the increasing depth of SNNs, the memory problem caused by the weights of SNNs has gradually attracted attention. In this study, we propose an ultra-low latency adaptive local binary spiking neural network (ALBSNN) with accuracy loss estimators, which dynamically selects the network layers to be binarized to ensure a balance between quantization degree and classification accuracy by evaluating the error caused by the binarized weights during the network learning process. At the same time, to accelerate the training speed of the network, the global average pooling (GAP) layer is introduced to replace the fully connected layers by combining convolution and pooling. Finally, to further reduce the error caused by the binary weight, we propose binary weight optimization (BWO), which updates the overall weight by directly adjusting the binary weight. This method further reduces the loss of the network that reaches the training bottleneck. The combination of the above methods balances the network's quantization and recognition ability, enabling the network to maintain the recognition capability equivalent to the full precision network and reduce the storage space by more than 20%. So, SNNs can use a small number of time steps to obtain better recognition accuracy. In the extreme case of using only a one-time step, we still can achieve 93.39, 92.12, and 69.55% testing accuracy on three traditional static datasets, Fashion- MNIST, CIFAR-10, and CIFAR-100, respectively. At the same time, we evaluate our method on neuromorphic N-MNIST, CIFAR10-DVS, and IBM DVS128 Gesture datasets and achieve advanced accuracy in SNN with binary weights. Our network has greater advantages in terms of storage resources and training time.

MADM-based smart parking guidance algorithm.

In smart parking environments, how to choose suitable parking facilities with various attributes to satisfy certain criteria is an important decision issue. Based on the multiple attributes decision making (MADM) theory, this study proposed a smart parking guidance algorithm by considering three representative decision factors (i.e., walk duration, parking fee, and the number of vacant parking spaces) and various preferences of drivers. In this paper, the expected number of vacant parking spaces is regarded as an important attribute to reflect the difficulty degree of finding available parking spaces, and a queueing theory-based theoretical method was proposed to estimate this expected number for candidate parking facilities with different capacities, arrival rates, and service rates. The effectiveness of the MADM-based parking guidance algorithm was investigated and compared with a blind search-based approach in comprehensive scenarios with various distributions of parking facilities, traffic intensities, and user preferences. Experimental results show that the proposed MADM-based algorithm is effective to choose suitable parking resources to satisfy users' preferences. Furthermore, it has also been observed that this newly proposed Markov Chain-based availability attribute is more effective to represent the availability of parking spaces than the arrival rate-based availability attribute proposed in existing research.

A quantitative theory and the generalized Bragg condition for surface plasmon Bragg reflectors.

We proposed a quantitative theory based on the surface plasmon polariton (SPP) coupled-mode model for SPP-Bragg reflectors composed of N periodic defects of any geometry and any refractive index profile. A SPP coupled-mode model and its recursive form were developed and shown to be equivalent. The SPP absorption loss, as well as high-order modes in each defect and possible radiation loss, is incorporated without effort. The simple recursive equations derived from the recursive model bridge the reflectance and the transmittance of N periodic defects to those of a single one, resulting in that the computational cost of the geometry optimization or the spectra calculation for N periodic defects is reduced into that for a single one. The model predictions show good agreement with fully vectorial computation data on the reflectance and the transmittance. From the recursive model, the generalized Bragg condition is proposed, which is verified by SPP-Bragg reflectors of various structures. The quantitative theory and the generalized Bragg condition proposed will greatly simplify the design of SPP-Bragg reflectors.