Single event effects evaluation on convolution neural network in Xilinx 28 nm system on chip

Convolutional neural networks(CNNs) exhibit excellent performance in the areas of image recognition and object detection, which can enhance the intelligence level of spacecraft. However, in aerospace, energetic particles, such as heavy ions, protons, and alpha particles, can induce single event effects(SEEs) that lead CNNs to malfunction and can significantly impact the reliability of a CNN system. In this paper, the MNIST CNN system was constructed based on a 28 nm systemon-chip(SoC), and then an alpha particle irradiation experiment and fault injection were applied to evaluate the SEE of the CNN system. Various types of soft errors in the CNN system have been detected, and the SEE cross sections have been calculated. Furthermore, the mechanisms behind some soft errors have been explained. This research will provide technical support for the design of radiation-resistant artificial intelligence chips.

ARTIFICIAL NEURAL NETWORK MODELLING OF A WOOD CHIP REFINER

1 INTRODUCTIONWood chip refining is the most critical step in mechanical pulping.Commercical experi-ences have been gained for years.Modelling and control of chip refiners,however,pose a challenge mainly because of the stochastic nature of the process.Some attemptshave been made to employ factor analysis technique[1]in the modelling andsimulating of refiner operation[2,3].Strand[2]used common factors as links betweenintrinsic fibre properties and pulp quality.He believed that a qualitative concept onthe physical nature of these common factors could be arrived at,and thus would helpto understand what refining variables need to be controlled or adjusted in order to im-prove pulp quality.However,the linear model used in factor analysis is based on theassumption that the interactions among the system variables are linear,which,ofcourse,is not true in practice.

Pattern recognition in multi-synaptic photonic spiking neural networks based on a DFB-SA chip

Spiking neural networks(SNNs)utilize brain-like spatiotemporal spike encoding for simulating brain functions.Photonic SNN offers an ultrahigh speed and power efficiency platform for implementing high-performance neuromorphic computing.Here,we proposed a multi-synaptic photonic SNN,combining the modified remote supervised learning with delayweight co-training to achieve pattern classification.The impact of multi-synaptic connections and the robustness of the network were investigated through numerical simulations.In addition,the collaborative computing of algorithm and hardware was demonstrated based on a fabricated integrated distributed feedback laser with a saturable absorber(DFB-SA),where 10 different noisy digital patterns were successfully classified.A functional photonic SNN that far exceeds the scale limit of hardware integration was achieved based on time-division multiplexing,demonstrating the capability of hardware-algorithm co-computation.

ANALYSIS AND DESIGN OF A NEURAL CHIP USED FOR BINARY IMAGE PROCESSING

Based on the model of a formal neuron proposed by McCulloch and Pitts,a kind ofneural circuit,which is a CMOS Variable Threshold Logic(VTL)circuit,is given in this paperconsidering the features of the binary image processing system.The theoretical analysis,andthe simulations for the building block circuits such as D/A converters,comparator and so on aregiven.The layout design of the whole circuit are also given.The binary image processing can berealized by using the VTL circuit combined with its external auxiliary circuits.

Efficient stochastic parallel gradient descent training for on-chip optical processor

In recent years,space-division multiplexing(SDM)technology,which involves transmitting data information on multiple parallel channels for efficient capacity scaling,has been widely used in fiber and free-space optical communication sys-tems.To enable flexible data management and cope with the mixing between different channels,the integrated reconfig-urable optical processor is used for optical switching and mitigating the channel crosstalk.However,efficient online train-ing becomes intricate and challenging,particularly when dealing with a significant number of channels.Here we use the stochastic parallel gradient descent(SPGD)algorithm to configure the integrated optical processor,which has less com-putation than the traditional gradient descent(GD)algorithm.We design and fabricate a 6×6 on-chip optical processor on silicon platform to implement optical switching and descrambling assisted by the online training with the SPDG algorithm.Moreover,we apply the on-chip processor configured by the SPGD algorithm to optical communications for optical switching and efficiently mitigating the channel crosstalk in SDM systems.In comparison with the traditional GD al-gorithm,it is found that the SPGD algorithm features better performance especially when the scale of matrix is large,which means it has the potential to optimize large-scale optical matrix computation acceleration chips.

A graph neural network approach to the inverse design for thermal transparency with periodic interparticle system

Recent years have witnessed significant advances in utilizing machine learning-based techniques for thermal metamaterial-based structures and devices to attain favorable thermal transport behaviors.Among the various thermal transport behaviors,achieving thermal transparency stands out as particularly desirable and intriguing.Our earlier work demonstrated the use of a thermal metamaterial-based periodic interparticle system as the underlying structure for manipulating thermal transport behavior and achieving thermal transparency.In this paper,we introduce an approach based on graph neural network to address the complex inverse design problem of determining the design parameters for a thermal metamaterial-based periodic interparticle system with the desired thermal transport behavior.Our work demonstrates that combining graph neural network modeling and inference is an effective approach for solving inverse design problems associated with attaining desirable thermal transport behaviors using thermal metamaterials.

Simulation and fabrication of in vitro microfluidic microelectrode array chip for patterned culture and electrophysiological detection of neurons

To enable the detection and modulation of modularized neural networks in vitro,this study proposes a microfluidic microelectrode array chip for the cultivation,compartmentalization,and control of neural cells.The chip was designed based on the specific structure of neurons and the requirements for detection and modulation.Finite-element analysis of the chip’s flow field was conducted using the COMSOL Multiphysics software,and the simulation results show that the liquid within the chip can flow smoothly,ensuring stable flow fields that facilitate the uniform growth of neurons within the microfluidic channels.By employing MEMS technology in combination with nanomaterial modification techniques,the microfluidic microelectrode array chip was fabricated successfully.Primary hippocampal neurons were cultured on the chip,forming a well-defined neural network.Spontaneous electrical activity of the detected neurons was recorded,exhibiting a 23.7%increase in amplitude compared to neuronal discharges detected on an open-field microelectrode array.This study provides a platform for the precise detection and modulation of patterned neuronal growth in vitro,potentially serving as a novel tool in neuroscience research.

小鼠NPC细胞RFX1ChIP-Seq数据分析

利用NCBI数据库中小鼠Embryonic Stem Cells(ESC)与Neural ProgenitorCells(NPC)的基因芯片结果及NPC时期的RFX1ChIP-Seq数据,进行有关RFX1的分析,结果表明:RFX1结合位点富集在1,2,4,5,7,9,11染色体上,Y染色体上最少,其他染色体上比较均衡;在基因组中结合位点分布区域主要在基因的promoter区域,约有53.2%,其次是intergentic,占22.5%,body区域,占13.1%,enhancer区域,占11.2%.说明RFX1是以结合在基因的promoter区为主要形式对目的基因进行调控.同时在DAVID数据库中用生物信息学方法探索了RFX1靶基因的生物学功能分类.

Cell cycle-related genes p57kip2, Cdk5 and Spin in the pathogenesis of neural tube defects

In the field of developmental neurobiology, accurate and ordered regulation of the cell cycle and apoptosis are crucial factors contributing to the normal formation of the neural tube. Preliminary studies identified several genes involved in the development of neural tube defects. In this study, we established a model of developmental neural tube defects by administration of retinoic acid to pregnant rats. Gene chip hybridization analysis showed that genes related to the cell cycle and apoptosis, signal transduction, transcription and translation regulation, energy and metabolism, heat shock, and matrix and cytoskeletal proteins were all involved in the formation of developmental neural tube defects. Among these, cell cycle-related genes were predominant. Retinoic acid treat-ment caused differential expression of three cell cycle-related genes p57kip2, Cdk5 and Spin, the expression levels of which were downregulated by retinoic acid and upregulated during normal neural tube formation. The results of this study indicate that cell cycle-related genes play an im-portant role in the formation of neural tube defects. P57kip2, Cdk5 and Spin may be critical genes in the pathogenesis of neural tube defects.

Analysis of hippocampal gene expression profile of Alzheimer's disease model rats using genome chip bioinformatics

In this study, an Alzheimer＇s disease model was established in rats through stereotactic injection of condensed amyloid beta 1-40 into the bilateral hippocampus, and the changes of gene expression profile in the hippocampus of rat models and sham-operated rats were compared by genome expression profiling analysis. Results showed that the expression of 50 genes was significantly up-regulated （fold change 〉 2）, while 21 genes were significantly down-regulated in the hippocampus of Alzheimer＇s disease model rats （fold change 〈 0.5） compared with the sham-operation group. The differentially expressed genes are involved in many functions, such as brain nerve system development, neuronal differentiation and functional regulation, cellular growth, differentiation and apoptosis, synaptogenesis and plasticity, inflammatory and immune responses, ion channels/transporters, signal transduction, cell material/energy metabolism. Our findings indicate that several genes were abnormally expressed in the metabolic and signal transduction pathways in the hippocampus of amyloid beta 1 40-induced rat model of Alzheimer＇s disease, thereby affecting the hippocampal and brain functions.

Weak cation exchange 2 protein chip for detecting differentially expressed brainstem proteins in a rat model of closed traumatic brain injury

BACKGROUND： Studies have reported the combined use of two-dimensional gel electrophoresis and mass spectrometry to detect differentially expressed proteins in the rat brainstem following brain injury. However, the detected differential proteins often exhibit low sensitivity and high relative molecular weight. Although protein chip technology is thought to compensate for these inadequacies, no related studies or results have been reported. OBJECTIVE： To propose the application of weak cation exchange protein chips in combination with mass spectrometry for determining protein expression profiles and characteristics in the brainstem following closed brain injury. DESIGN, TIME AND SETTING： Randomized, controlled, animal experiments utilizing proteomics were performed from June 2007 to December 2008 in the Proteomics Laboratory, Medical College of Chinese People＇s Armed Police Force. MATERIALS： Weak cation exchange 2 protein chip, Ciphergen Proteinchip System （PBS-IIC）. METHODS： A total of 72 rats were randomly assigned to two groups： sham-surgery （n = 12） and injury （n = 60）. A closed traumatic brain injury model caused by falling object was replicated in the injury group, which was then subdivided into five subgroups according to different time points after injury： 4, 8, 12, 24, and 48 hours, with 12 rats in each subgroup. In the sham-surgery group, only the skin was removed and the stainless steel pad was fixed to the skull. MAIN OUTCOME MEASURES： The brain injury rats were sacrificed at 4, 8, 12, 24, and 48 hours after injury, respectively, and the control rats were sacrificed at 24 hours. Pathological changes in the brainstem were determined using hematoxylin-eosin staining, and differential protein expression in the brainstem was detected using a weak cation exchange 2 protein chip and protein chip reader. RESULTS： In the sham-surgery group, cells appeared normal. However, in the brain injury group, some brainstem neurons exhibited pyknosis, with reduced numbers of Nissl bodies in the cytoplasm swollen cell bodies and nuclei, irregular staining in the cytoplasm, and decreased numbers of neurons. Results from weak cation exchange 2 protein chip detection demonstrated that, compared with the sham-surgery group, the expression profiles of 2 proteins were altered in the brainstem of the injury group. At 12, 24, and 48 hours after injury, expression was increased （P 〈 0.01 ）. The mass charge ratio （M/Z） of 7 862 differentially expressed proteins was greater in the sham-surgery group compared with 12 and 24 hours after injury （P 〈 0.05）. CONCLUSION： The combined method of weak cation exchange 2 protein chip and mass spectrometry detected differential protein expression in the brainstem following closed brain injury in the rats, which suggested that closed brain injury induced altered protein expression profiles in the brainstem.

On－Line Identification of Grinding Burn and Wheel Wear Based on Grinding Chips Thermal Flow

The theoretical basis of the grinding chips thermal flow being regarded as the characteristic signal of on line identification is summarized. And on line identification of grinding burn and wheel wear based on the grinding chips thermal flow is introduc

A Miniaturized System for Neural Signal Acquiring and Processing

To collect neural activity data from awake, behaving freely animals, we develop miniaturized implantable recording system by the modem chip：Programmable System on Chip （PSoC） and through chronic electrodes in the cortex. With PSoC family member CY8C29466,the system completed operational and instrument amplifiers, filters, timers, AD convertors, and serial communication, etc. The signal processing was dealt with virtual instrument technology. All of these factors can significantly affect the price and development cycle of the project. The result showed that the system was able to record and analyze neural extrocellular discharge generated by neurons continuously for a week or more. This is very useful for the interdisciplinary research of neuroscience and information engineering technique. The circuits and architecture of the devices can be adapted for neurobiology and research with other small animals.

Cooperative Content Caching and Delivery in Vehicular Networks: A Deep Neural Network Approach

The growing demand for low delay vehicular content has put tremendous strain on the backbone network.As a promising alternative,cooperative content caching among different cache nodes can reduce content access delay.However,heterogeneous cache nodes have different communication modes and limited caching capacities.In addition,the high mobility of vehicles renders the more complicated caching environment.Therefore,performing efficient cooperative caching becomes a key issue.In this paper,we propose a cross-tier cooperative caching architecture for all contents,which allows the distributed cache nodes to cooperate.Then,we devise the communication link and content caching model to facilitate timely content delivery.Aiming at minimizing transmission delay and cache cost,an optimization problem is formulated.Furthermore,we use a multi-agent deep reinforcement learning(MADRL)approach to model the decision-making process for caching among heterogeneous cache nodes,where each agent interacts with the environment collectively,receives observations yet a common reward,and learns its own optimal policy.Extensive simulations validate that the MADRL approach can enhance hit ratio while reducing transmission delay and cache cost.

Design of 3D Active Multichannel Silicon Neural Microelectrode

To find a design method for 3D active multichannel silicon microelectrode, a microstructure of active neural recording system is presented, where two 2D probes, two integrated circuits and two spacers are microassembled on a 5 mm×7 mm silicon platform, and 32 sites neural signals can be operated simultaneously. A theoretical model for measuring the neural signal by the silicon microelectrode is proposed based on the structure and fabrication process of a single-shank probe. The method of determining the dimensional parameters of the probe shank is discussed in the following three aspects, i.e. the structures of pallium and endocranium, coupled interconnecters noise, and strength characteristic of neural probe. The design criterion is to minimize the size of the neural probe as well as that the probe has enough stiffness to pierce the endocranium. The on-chip unity-gain bandpass amplifier has an overall gain of 42 dB over a bandwidth from 60 Hz to 10 kHz; and the DC-baseline stability circuit is of high input resistance above 30 MΩ to guarantee a cutoff frequency below 100 Hz. The circuit works in stimulating or recording modes. The conversion of the modes depends on the stimulating control signal.

Residential Community Open-Up Strategy Based on Prim’s Algorithm and Neural Network Algorithm

“Open community” has aroused widespread concern and research. This paper focuses on the system analysis research of the problem that based on statistics including the regression equation fitting function and mathematical theory, combined with the actual effect of camera measurement method, Prim’s algorithm and neural network to “Open community” and the applicable conditions. Research results show that with the increasing number of roads within the district, the benefit time gradually increased, but each type of district capacity is different.

THE CONTROL, DRIVE AND DISPLAY CIRCUITS ANDDEVICES IN OPTO-ELECTRONIC 2-D PROGRAMMABLE NEURAL NETWORK

In an optoelectronic 2-D programmable neural network system, optical data need to be transferred and feedback with high speed. This paper presents the design and implementation of the interface circuit and its software.

Design of an Automated Sorting System for Apples Based on Single Chip Microcomputer

The grading judgment for apples is related to a variety of factors including,size,shape,color,texture,and scars.Traditional manual sorting methods are time consuming and labor intensive.In addition,the accuracy of the method is easily subjective,not repeatable,error-prone,and affected by the sorting environment.This paper presents a complete and automated grading system for apples.The system uses a single-chip microcomputer as the controller of the system,and a PC as the graphics processing unit.It also includes a conveyor,drive motor,frequency converter for motor control,photoelectric sensors,air compressor,and air jets for ejecting the graded apples.The classification algorithm is implemented by using a convolutional neural network(CNN).In order to eliminate contact damage of apples,the system specifically uses air jets as actuators to eject the graded apples into the corresponding bins.At the same time,in order to ensure that an apple triggers the correct ejecting actuator,this paper designs a jet controller with proper logic.

光子神经网络研究进展

在数据海量化、信息化的时代,电子计算机处理系统所面临的算力和能耗等性能要求愈发严苛,传统冯·诺依曼架构存在“内存墙”和“功耗墙”瓶颈,加之摩尔定律放缓甚至失效,使得电子芯片在计算速度和功耗方面遇到极大挑战,利用光计算替代传统电子计算将是解决当前算力与功耗问题的极具潜力的途径之一。本文系统地梳理了片上集成和自由空间的光子神经网络架构与算法方面的研究进展,详细介绍了典型的研究工作,然后讨论并对比了这两种光子神经网络的优劣势,以及光子神经网络的训练策略等。最后探讨了光子神经网络面临的挑战,并对其未来发展进行了前瞻性的展望。

基于RT-YOLO-V5的芯片外观缺陷检测

针对传统的人工芯片检测方法效率低、过分依赖人为操作且误检率高等产生的问题,提出了一种基于ResCBS模块与增加微检测层(Tiny-scale detection layer)的RT-YOLO-V5检测方法用于检测芯片外观缺陷。首先搭建了图像采集系统,并制作了芯片外观缺陷检测数据集。为解决芯片外观缺陷形状不规则、大小不统一、位置不确定带来的检测精度低等问题,在CBS模块中增加短连接,融合输入输出的特征信息,减少信息损失,优化推理速度;其次,增加一个微小尺度的检测层,提高模型对微小目标的特征提取能力。实验结果表明:使用改进后的网络对芯片外观缺陷进行检测,平均精度(mAP)达到95.5%,相对于原始网络提升了5.7%;除此之外,改进后的RT-YOLO-V5在先验框损失(Box_loss)与小目标缺陷的检测精度上都得到了一定的提升。