A Design Methodology for Fault-Tolerant Neuromorphic Computing Using Bayesian Neural Network.

Memristor crossbar arrays are a promising platform for neuromorphic computing. In practical scenarios, the synapse weights represented by the memristors for the underlying system are subject to process variations, in which the programmed weight when read out for inference is no longer deterministic but a stochastic distribution. It is therefore highly desired to learn the weight distribution accounting for process variations, to ensure the same inference performance in memristor crossbar arrays as the design value. In this paper, we introduce a design methodology for fault-tolerant neuromorphic computing using a Bayesian neural network, which combines the variational Bayesian inference technique with a fault-aware variational posterior distribution. The proposed framework based on Bayesian inference incorporates the impacts of memristor deviations into algorithmic training, where the weight distributions of neural networks are optimized to accommodate uncertainties and minimize inference degradation. The experimental results confirm the capability of the proposed methodology to tolerate both process variations and noise, while achieving more robust computing in memristor crossbar arrays.

Efficient Layer-Wise N:M Sparse CNN Accelerator with Flexible SPEC: Sparse Processing Element Clusters.

Recently, the layer-wise N:M fine-grained sparse neural network algorithm (i.e., every M-weights contains N non-zero values) has attracted tremendous attention, as it can effectively reduce the computational complexity with negligible accuracy loss. However, the speed-up potential of this algorithm will not be fully exploited if the right hardware support is lacking. In this work, we design an efficient accelerator for the N:M sparse convolutional neural networks (CNNs) with layer-wise sparse patterns. First, we analyze the performances of different processing element (PE) structures and extensions to construct the flexible PE architecture. Second, the variable sparse convolutional dimensions and sparse ratios are involved in the hardware design. With a sparse PE cluster (SPEC) design, the hardware can efficiently accelerate CNNs with the layer-wise N:M pattern. Finally, we employ the proposed SPEC into the CNN accelerator with flexible network-on-chip and specially designed dataflow. We implement hardware accelerators on Xilinx ZCU102 FPGA and Xilinx VCU118 FPGA and evaluate them with classical CNNs such as Alexnet, VGG-16, and ResNet-50. Compared with existing accelerators designed for structured and unstructured pruned networks, our design achieves the best performance in terms of power efficiency.

Motivation-based segmentation of game meat consumers: A look at the beliefs of food consumers during the COVID-19 crisis in China.

OBJECTIVES: This research aims to explore the factors motivate consumers to eat game meat during a multi-state disease outbreak. METHODS: It proposes a segmentation of consumers based on their attitudes toward and reveals the consumers' food beliefs that motivate their actions. Three segments of game meat consumers were identified: identity seekers, health seekers, and taste seekers. RESULTS: A survey of the potential impact that the COVID-19 crisis has on these three clusters' future food choices showed that the identity and health seekers are more open to a change in food choices. However, the taste seekers are less likely to be influenced by external factors. CONCLUSIONS: This research indicates that for the policymakers, the key is to take game meat consumers as an effective intervention entry point. It is crucial to facilitate healthy food choices and to promote socially- and culturally-appropriate food beliefs by improving public awareness of the risks of game meat, and invest in organic food. RESEARCH IMPLICATIONS: This research provides new insights into the food beliefs of game meat consumers via motivation-based segmentation.

Generational Differences in Perceptions of Food Health/Risk and Attitudes toward Organic Food and Game Meat: The Case of the COVID-19 Crisis in China.

In December 2019, a novel laboratory-confirmed coronavirus (2019-nCoV) infection, which has caused clusters of severe illnesses, was first reported in Wuhan, the capital of Hubei province, China. This foodborne illness, which reportedly most likely originated in a seafood market where wild animals are sold illegally, has transmitted among humans through close contact, across the world. The aim of this study is to explore health/risk perceptions of and attitudes toward healthy/risky food in the immediate context of food crisis. More specifically, by using the data collected from 1008 respondents in January 2020, the time when China was hit hard by the "Corona Virus Disease 2019" (COVID-19), this study investigates the overall and different generational respondents' health/risk perceptions of and attitudes toward organic food and game meat. The results reveal that, firstly, based on their food health and risk perceptions of healthy and risky food, the respondents' general attitudes are positive toward organic food but relatively negative toward game meat. Secondly, older generations have a more positive attitude and are more committed to organic food. Younger generations' attitude toward game meat is more negative whereas older generations attach more importance to it because of its nutritional and medicinal values. In addition, this research also indicates that the COVID-19 crisis influences the respondents' perceptions of and attitudes toward organic food and game meat consumption. However, the likelihood of its impact on older generations' future change in diets is smaller, which implies that older generations' food beliefs are more stable.