Multi-Level Parallel Network for Brain Tumor Segmentation

Accurate automatic segmentation of gliomas in various sub-regions,including peritumoral edema,necrotic core,and enhancing and non-enhancing tumor core from 3D multimodal MRI images,is challenging because of its highly heterogeneous appearance and shape.Deep convolution neural networks(CNNs)have recently improved glioma segmentation performance.However,extensive down-sampling such as pooling or stridden convolution in CNNs significantly decreases the initial image resolution,resulting in the loss of accurate spatial and object parts information,especially information on the small sub-region tumors,affecting segmentation performance.Hence,this paper proposes a novel multi-level parallel network comprising three different level parallel subnetworks to fully use low-level,mid-level,and high-level information and improve the performance of brain tumor segmentation.We also introduce the Combo loss function to address input class imbalance and false positives and negatives imbalance in deep learning.The proposed method is trained and validated on the BraTS 2020 training and validation dataset.On the validation dataset,ourmethod achieved a mean Dice score of 0.907,0.830,and 0.787 for the whole tumor,tumor core,and enhancing tumor core,respectively.Compared with state-of-the-art methods,the multi-level parallel network has achieved competitive results on the validation dataset.

WiFi CSI Gesture Recognition Based on Parallel LSTM-FCN Deep Space-Time Neural Network

In this study,we developed a system based on deep space–time neural networks for gesture recognition.When users change or the number of gesture categories increases,the accuracy of gesture recognition decreases considerably because most gesture recognition systems cannot accommodate both user differentiation and gesture diversity.To overcome the limitations of existing methods,we designed a onedimensional parallel long short-term memory–fully convolutional network(LSTM–FCN)model to extract gesture features of different dimensions.LSTM can learn complex time dynamic information,whereas FCN can predict gestures efficiently by extracting the deep,abstract features of gestures in the spatial dimension.In the experiment,50 types of gestures of five users were collected and evaluated.The experimental results demonstrate the effectiveness of this system and robustness to various gestures and individual changes.Statistical analysis of the recognition results indicated that an average accuracy of approximately 98.9% was achieved.

DIRECT DISPLACEMENT OF PARALLEL MECHANISM WITH WAVELET NETWORK

A new method solution for the direct displacement of parallel mechanism, wavelet network method, is proposed. Comparing with the classical analytical and numerical methods, this method can be extended to any parallel mechanism with any selected degree of freedom and configuration. A wavelet network suiting to approach multi-input and multi-output system is constructed. The network is optimized by analyzing the sparseness of input data and selecting the fitting wavelets by orthogonalization method according to the output data. Then it is applied to solve the direct displace- ment of a general six-degree-of-freedom parallel mechanism as a numerical example. For comparison purposes, a BP neural network is also used for this problem. Simulation results show that the wavelet network performs better than BP neural network. In addition, the wavelet network learns much faster than BP network.

Parallel Neural Network-Based Motion Controller for Autonomous Underwater Vehicles

A parallel neural network-based controller （PNNC） is presented for the motion control of underwater vehicles in this paper. It consists of a real-time part, a self-learning part and a desired-state programmer, and it is different from normal adaptive neural network controller in structure. Owing to the introduction of the self-learning part, on-line learning can be performed without sample data in several sample periods, resulting in high learning speed of the controller and good control performance. The desired-state programmer is utilized to obtain better learning samples of the neural network to keep the stability of the controller. The developed controller is applied to the 4-degree of freedom control of the AUV “IUV- IV” and is successful on the simulation platform. The control performance is also compared with that of neural network controller with different structures such as normal adaptive neural network and different learning methods. Current effects and surge velocity control are also included to demonstrate the controller＇ s performance. It is shown that the PNNC has a great possibility to solve the problems in the control system design of underwater vehicles.

Towards efficient deep neural network training by FPGA-based batch-level parallelism

Training deep neural networks(DNNs)requires a significant amount of time and resources to obtain acceptable results,which severely limits its deployment in resource-limited platforms.This paper proposes DarkFPGA,a novel customizable framework to efficiently accelerate the entire DNN training on a single FPGA platform.First,we explore batch-level parallelism to enable efficient FPGA-based DNN training.Second,we devise a novel hardware architecture optimised by a batch-oriented data pattern and tiling techniques to effectively exploit parallelism.Moreover,an analytical model is developed to determine the optimal design parameters for the DarkFPGA accelerator with respect to a specific network specification and FPGA resource constraints.Our results show that the accelerator is able to perform about 10 times faster than CPU training and about a third of the energy consumption than GPU training using 8-bit integers for training VGG-like networks on the CIFAR dataset for the Maxeler MAX5 platform.

Designing and optimizing a parallel neural network model for predicting the solubility of diosgenin in n-alkanols

Accurate estimation of the solubility of a chemical compound is an important issue for many industrial proce sses.To overcome the defects of some thermodynamic models and simple correlations,a parallel neural network(PNN) model was conceived and optimized to predict the solubility of diosgenin in seven n-alkanols(C_(1)-C_(7)).The linear regression analysis of the parity plots indicates that the PNN model can give more accurate descriptions of the solubility of diosgenin than the ordinary neural network(ONN) model.The comparison of the average root mean square deviation(RMSD) shows that the suggested model has a slight advantage over the thermodynamic NRTL model in terms of the calculating precision.Moreover,the PNN model can reflect the effects of the temperature and the chain length of the alcohol solvent on the solution behavior of diosgenin correctly and can estimate its solubility in the n-alkanols with more carbon atoms.

Multilayer Hex-Cells: A New Class of Hex-Cell Interconnection Networks for Massively Parallel Systems

Scalability is an important issue in the design of interconnection networks for massively parallel systems. In this paper a scalable class of interconnection network of Hex-Cell for massively parallel systems is introduced. It is called Multilayer Hex-Cell (MLH). A node addressing scheme and routing algorithm are also presented and discussed. An interesting feature of the proposed MLH is that it maintains a constant network degree regardless of the increase in the network size degree which facilitates modularity in building blocks of scalable systems. The new addressing node scheme makes the proposed routing algorithm simple and efficient in terms of that it needs a minimum number of calculations to reach the destination node. Moreover, the diameter of the proposed MLH is less than Hex-Cell network.

Adaptive control of parallel manipulators via fuzzy-neural network algorithm

This paper considers adaptive control of parallel manipulators combined with fuzzy-neural network algorithms （FNNA）. With this algorithm, the robustness is guaranteed by the adaptive control law and the parametric uncertainties are eliminated. FNNA is used to handle model uncertainties and external disturbances. In the proposed control scheme, we consider modifying the weight of fuzzy rules and present these rules to a MIMO system of parallel manipulators with more than three degrees-of-freedom （DoF）. The algorithm has the advantage of not requiring the inverse of the Jacobian matrix especially for the low DoF parallel manipulators. The validity of the control scheme is shown through numerical simulations of a 6-RPS parallel manipulator with three DoF.

Parallel-Transmission: A New Usage of Multi-Radio Diversity in Wireless Mesh Network

To fully utilize the diversity of multi-radio, a new parallel transmission method for wireless mesh network is proposed. Compared with conventional packet transmission which follows “one flow on one radio”, it uses the radio diversity to transmit the packets on different radios simultaneously. Three components are presented to achieve parallel-transmission, which are control module, selection module and schedule module. A localized selecting algorithm selects the right radios based on the quality of wireless links. Two kinds of distributed scheduling algorithms are implemented to transmit packets on the selected radios. Finally, a parallel-adaptive routing metric is presented. Simulation results by NS2 show that this parallel-transmission scheme could enhance the average throughput of network by more than 10%.

NEURAL NETWORK TRAINING WITH PARALLEL PARTICLE SWARM OPTIMIZER

Objective To reduce the execution time of neural network training. Methods Parallel particle swarm optimization algorithm based on master-slave model is proposed to train radial basis function neural networks, which is implemented on a cluster using MPI libraries for inter-process communication. Results High speed-up factor is achieved and execution time is reduced greatly. On the other hand, the resulting neural network has good classification accuracy not only on training sets but also on test sets. Conclusion Since the fitness evaluation is intensive, parallel particle swarm optimization shows great advantages to speed up neural network training.

基于混合分解和PCG-BiLSTM的风速短期预测

为降低风速的随机性对风力发电的影响,提高风速短期预测的精准度,提出一种基于混合分解、双通道输入、多分支PCG-BiLSTM深度学习模型的短期风速预测方法。首先,将全年风速数据分为春、夏、秋、冬4个季度,选取春季作为主要实验对象;其次,利用奇异谱分解(SSD)和变分模态分解(VMD)以降低原始春季风速数据复杂度,生成具有不同模态且复杂度低的子分量,两种不同模式子分量组合为混合分量,实现不同模式分解算法的优势互补;最后,将混合分量以双通道的形式输入到多分支PCG-BiLSTM深度学习模型中,其模型的每个分支由卷积神经网络(CNN)与门控循环单元(GRU)并联组成时空特征提取模块,用于提取两种分解分量组合的混合分量的时空特征,各分支提取对应混合分量的时空特征经聚合后再由双向长短期记忆网络(BiLSTM)进一步提取风速信号的正向和反向双向波动规律,进而得到最终的风速预测结果。多组实验结果表明:提出的组合预测方法在短期风速预测中具有较高的精度和泛化能力,优于其他传统预测方法。

A Coupled Transiently Chaotic Neural Network Approach for Identical Parallel Machine Scheduling

在相同机器上安排工作是经常在各种各样的生产系统遇到的一种状况。在这份报纸，一新联合了短暂地混乱的神经网络(CTCNN ) 被提出解决相同平行机器安排。这个问题的一个混合整数编程模型被介绍一个排列矩阵表达式转变成 CTCNN 计算体系结构。新计算精力功能被建议除所有限制以外表示目的。特别地，在精力功能在惩罚术语之中存在的折衷问题被使用变化时间的惩罚参数克服。最后，结果与 100 个随机的起始的条件在 3 个不同规模问题上测试了证明网络收敛并且能在合理时间解决这些问题。

基于VIKOR的多网并行传输选网算法

海上异构无线网络通信场景下,针对单一无线网络难以承载大带宽业务的现实问题,提出了一种基于VIKOR的多网并行传输网络选择算法。该算法首先将多个网络的参数聚合,筛选可用网络组合;然后,结合网络属性的主客观权重和用户偏好权重得到网络属性的综合权重;最后,根据VIKOR方法对网络组合进行排序,选出最佳网络组合方案。仿真结果表明:该算法具有可行性,相较于传统算法,在舰船通信业务繁忙情况下选网综合性能更优。

长向量处理器高效RNN推理方法

模型深度的不断增加和处理序列长度的不一致对循环神经网络在不同处理器上的性能优化提出巨大挑战。针对自主研制的长向量处理器FT-M7032,实现了一个高效的循环神经网络加速引擎。该引擎采用行优先矩阵向量乘算法和数据感知的多核并行方式,提高矩阵向量乘的计算效率;采用两级内核融合优化方法降低临时数据传输的开销;采用手写汇编优化多种算子,进一步挖掘长向量处理器的性能潜力。实验表明,长向量处理器循环神经网络推理引擎可获得较高性能,相较于多核ARM CPU以及Intel Golden CPU,类循环神经网络模型长短记忆网络可获得最高62.68倍和3.12倍的性能加速。

Parallel and optimized genetic Elman network for ^(252)Cf source-driven verification system

The 252Cf source-driven verification system(SDVS)can recognize the enrichment of fissile material with the enrichment-sensitive autocorrelation functions of a detector signal in252Cf source-driven noise-analysis(SDNA)measurements.We propose a parallel and optimized genetic Elman network(POGEN)to identify the enrichment of235U based on the physical properties of the measured autocorrelation functions.Theoretical analysis and experimental results indicate that,for 4 different enrichment fissile materials,due to higher information utilization,more efficient network architecture,and optimized parameters,the POGEN-based algorithm can obtain identification results with higher recognition accuracy,compared to the integrated autocorrelation function(IAF)method.

面向微控制器的卷积神经网络加速器设计

针对目前嵌入式微控制器的性能难以满足实时图像识别任务的问题,提出一种适用于微控制器的卷积神经网络加速器。该加速器在卷积层设计了无阻塞的行并行乘法-加法树结构,获得了更高的硬件利用率;为了满足行并行的数据吞吐量,设计了卷积专用SRAM存储器。加速器将池化和激活单元融入数据通路,有效减少数据重复存取带来的时间开销。FPGA原型验证表明加速器的性能达到92.2 GOPS@100 MHz;基于TSMC 130 nm工艺节点进行逻辑综合,加速器的动态功耗为33 mW,面积为90 764.2μm^(2),能效比高达2 793 GOPS/W,比FPGA加速器方案提高了约100倍。该加速器低功耗、低成本的特性,有利于实现嵌入式系统在目标检测、人脸识别等机器视觉领域的广泛应用。

A Highly Compatible Circular-Shifting Network for Partially Parallel QC-LDPC Decoder

The conventional methodology for designing QC-LDPC decoders is applied for fixed configurations used in wireless communication standards, and the supported largest expansion factor Z (the parallelism of the layered decoding) is a fixed number. In this paper, we study the circular-shifting network for decoding LDPC codes with arbitrary Z factor, especially for decoding large Z (Z P) codes, where P is the decoder parallelism. By buffering the P-length slices from the memory, and assembling the shifted slices in a fixed routine, the P-parallelism shift network can process Z-parallelism circular-shifting tasks. The implementation results show that the proposed network for arbitrary sized data shifting consumes only one times of additional resource cost compared to the traditional solution for only maximum P sized data shifting, and achieves significant saving on area and routing complexity.

基于坐标信息与多尺度并行网络的气道分割方法

为解决手术导航中气道模型精度不足的问题,提出了一种基于坐标信息与多尺度并行网络的气道分割方法。首先通过并行网络分别学习不同尺度的气道特征,以解决不同尺寸气道之间的特征冲突问题。其次提出坐标引导的上采样模块,通过浅层特征中的坐标信息指导深层特征进行特征重建,限制目标的空间位置,提高模型精度。最后提出通道引导的多尺度特征聚合模块,用于在多个尺度上捕获语义信息并探索不同尺度特征之间的通道关系。在公开数据集LIDC-IDRI和EXACT'09上对提出的方法和其他模型进行训练和测试。实验表明,该方法的平均骰子系数达到了93.20%,相比于3D U-Net提高了2.61%,而假阳性率只有0.012%。此外,树长检测率和分支检测率分别达到了88.59%和97.42%。该方法可用于肺部疾病诊断或导航支气管检查等领域。

新集二矿承压水上开采底板破坏深度研究

为提高新集二矿1煤回采上限以及为水害防治提供依据,对新集二矿地质资料充分分析,采用UDEC 7.0进行数值模拟试验,模拟工作面推进时底板发育情况,得出工作面推进至120 m,底板达到最大破坏深度19.3 m。采用并行电法对煤层地板岩层破裂过程进行现场监测,通过观察电阻率变化趋势,判断岩体的形变过程,确定煤层采动破坏带深度为19.6 m。结果表明,数值模拟结果与电法测试结果基本一致,底板破坏带以下为底板岩层保护带,保护带以下岩层完整性较好,可以起到一定的隔水作用。研究成果对相似地质条件下的承压水上开采研究具有重要意义。