Trainable Spiking-YOLO for low-latency and high-performance object detection.

Spiking neural networks (SNNs) are considered an attractive option for edge-side applications due to their sparse, asynchronous and event-driven characteristics. However, the application of SNNs to object detection tasks faces challenges in achieving good detection accuracy and high detection speed. To overcome the aforementioned challenges, we propose an end-to-end Trainable Spiking-YOLO (Tr-Spiking-YOLO) for low-latency and high-performance object detection. We evaluate our model on not only frame-based PASCAL VOC dataset but also event-based GEN1 Automotive Detection dataset, and investigate the impacts of different decoding methods on detection performance. The experimental results show that our model achieves competitive/better performance in terms of accuracy, latency and energy consumption compared to similar artificial neural network (ANN) and conversion-based SNN object detection model. Furthermore, when deployed on an edge device, our model achieves a processing speed of approximately from 14 to 39 FPS while maintaining a desirable mean Average Precision (mAP), which is capable of real-time detection on resource-constrained platforms.

Effects of mobile phone App-based continuing nursing care on self-efficacy, quality of life, and motor function of stroke patients in the community.

To explore the effects of mobile phone application (App)-based continuing nursing care on the self-efficacy (SE), quality of life (QOF), and motor function (MF) of stroke patients in the community. A total of 101 stroke patients in the community recruited in this study for retrospective analysis were divided into a control group (CG) and an observation group (OG) based on the means of intervention. In total, 50 patients in the CG received routine community health education, based on which a mobile phone App-based continuing nursing mode was applied to the 51 patients in the OG. Changes in physiological indicators, including homocysteine (Hcy), high-density lipoprotein (HDL-C), and total cholesterol (TC), were evaluated before and after intervention. Moreover, MF [determined using the Fugal-Meyer motor function assessment (FMA)], SE (determined using stroke self-efficacy questionnaire), QOF, and satisfaction toward nursing were evaluated. (1) Hcy and TC levels in the OG were lower after intervention; however, HDL-C levels were higher than those in the CG, with statistically significant differences (P < 0.05). (2) The FMA MF of the upper and lower limb (FMA-U and FMA-L) scores and the total scores in the OG after the intervention were significantly improved compared with those in the CG (P < 0.05). (3) Patients in the OG showed significantly higher SE scores than those in the CG (P < 0.05). (4) Scores of emotional health, emotional function, social function, energy, general health status, body pain, physiological function, and physiological features were significantly higher in the OG than those in the CG after the intervention (P < 0.05). (5) Patients in the OG expressed more positive satisfaction toward nursing than those in the CG, with statistically significant differences (P < 0.05). Mobile phone App-based continuing nursing care may significantly improve the SE, quality of life, and satisfaction toward nursing as well as promote the improvement of biological markers and MF of stroke patients.

Reinforcement Learning in Spiking Neural Networks with Stochastic and Deterministic Synapses.

Though succeeding in solving various learning tasks, most existing reinforcement learning (RL) models have failed to take into account the complexity of synaptic plasticity in the neural system. Models implementing reinforcement learning with spiking neurons involve only a single plasticity mechanism. Here, we propose a neural realistic reinforcement learning model that coordinates the plasticities of two types of synapses: stochastic and deterministic. The plasticity of the stochastic synapse is achieved by the hedonistic rule through modulating the release probability of synaptic neurotransmitter, while the plasticity of the deterministic synapse is achieved by a variant of a reward-modulated spike-timing-dependent plasticity rule through modulating the synaptic strengths. We evaluate the proposed learning model on two benchmark tasks: learning a logic gate function and the 19-state random walk problem. Experimental results show that the coordination of diverse synaptic plasticities can make the RL model learn in a rapid and stable form.