In order to study fracture mechanism of rocks in different brittle mineral contents,this study pro-poses a method to identify the acoustic emission signal released by rock fracture under different brittle miner-al con...In order to study fracture mechanism of rocks in different brittle mineral contents,this study pro-poses a method to identify the acoustic emission signal released by rock fracture under different brittle miner-al content(BMC),and then determine the content of brittle matter in rock.To understand related interference such as the noises in the acoustic emission signals released by the rock mass rupture,a 1DCNN-BLSTM network model with SE module is constructed in this study.The signal data is processed through the 1DCNN and BLSTM networks to fully extract the time-series correlation features of the signals,the non-correlated features of the local space and the weak periodicity law.Furthermore,the processed signals data is input into the fully connected layers.Finally,softmax function is used to accurately identify the acoustic emission signals released by different rocks,and then determine the content of brittle minerals contained in rocks.Through experimental comparison and analysis,1DCNN-BLSTM model embedded with SE module has good anti-noise performance,and the recognition accuracy can reach more than 90 percent,which is better than the traditional deep network models and provides a new way of thinking for rock acoustic emission re-search.展开更多
针对目前最先进的临床命名实体识别(Cinical Named Entity Recognition,CNER)模型未能充分挖掘文本的全局信息和语义特征,以及未能解决文本中的字符替换等问题,改进了传统的单词嵌入模型,并在此基础上提出了一种结合深度卷积神经网络和...针对目前最先进的临床命名实体识别(Cinical Named Entity Recognition,CNER)模型未能充分挖掘文本的全局信息和语义特征,以及未能解决文本中的字符替换等问题,改进了传统的单词嵌入模型,并在此基础上提出了一种结合深度卷积神经网络和双向短时记忆条件随机场(DCNN-BiLSTM-CRF)的临床文本命名实体识别方法。改进的单词嵌入模型融合词根、拼音和字符本身意义,使用了来自Transformers的双向编码器表示,使单词嵌入向量具有汉字和临床文本的特点,该方法通过在临床命名实体识别任务中引入深度卷积神经网络(Deep Convolutional Neural Networks,DCNN),解决了CNN预测时丢失部分信息无法找回的问题。通过使用DCNN,本文模型能够更有效地捕获全局信息、获取字符之间的权重关系和多层次语义特征信息,从而提高了临床命名实体识别的准确性。在数据集CCKS2017和CCKS2018上分别进行实验,实验结果表明,与基准模型相比,该模型F1值分别改善了0.48%,0.68%,0.6%,0.58%,0.04%和1.43%,2.36%,3.31%,1.11%,0.17%。为了进一步验证本文的模型,进行了两种消融实验。结果表明,在两个数据集CCKS2017和CCKS2018上本文模型对比变体模型M1,F1值分别改善了0.79%和0.84%;对比变体模型M2,F1值分别改善了0.53%和0.64%。这些实验结果证明了本文所提算法的可行性。展开更多
With the continuous development of deep learning,Deep Convolutional Neural Network(DCNN)has attracted wide attention in the industry due to its high accuracy in image classification.Compared with other DCNN hard-ware ...With the continuous development of deep learning,Deep Convolutional Neural Network(DCNN)has attracted wide attention in the industry due to its high accuracy in image classification.Compared with other DCNN hard-ware deployment platforms,Field Programmable Gate Array(FPGA)has the advantages of being programmable,low power consumption,parallelism,and low cost.However,the enormous amount of calculation of DCNN and the limited logic capacity of FPGA restrict the energy efficiency of the DCNN accelerator.The traditional sequential sliding window method can improve the throughput of the DCNN accelerator by data multiplexing,but this method’s data multiplexing rate is low because it repeatedly reads the data between rows.This paper proposes a fast data readout strategy via the circular sliding window data reading method,it can improve the multiplexing rate of data between rows by optimizing the memory access order of input data.In addition,the multiplication bit width of the DCNN accelerator is much smaller than that of the Digital Signal Processing(DSP)on the FPGA,which means that there will be a waste of resources if a multiplication uses a single DSP.A multiplier sharing strategy is proposed,the multiplier of the accelerator is customized so that a single DSP block can complete multiple groups of 4,6,and 8-bit signed multiplication in parallel.Finally,based on two strategies of appeal,an FPGA optimized accelerator is proposed.The accelerator is customized by Verilog language and deployed on Xilinx VCU118.When the accelerator recognizes the CIRFAR-10 dataset,its energy efficiency is 39.98 GOPS/W,which provides 1.73×speedup energy efficiency over previous DCNN FPGA accelerators.When the accelerator recognizes the IMAGENET dataset,its energy efficiency is 41.12 GOPS/W,which shows 1.28×−3.14×energy efficiency compared with others.展开更多
基金Supported by projects of the National Natural Science Foundation of China(Nos.52074088,52174022,51574088,51404073)Provincial Outstanding Youth Reserve Talent Project of Northeast Petroleum University(No.SJQH202002)+1 种基金2020 Northeast Petroleum University Western Oilfield Development Special Project(No.XBYTKT202001)Postdoctoral Research Start-Up in Heilongjiang Province(Nos.LBH-Q20074,LBH-Q21086).
文摘In order to study fracture mechanism of rocks in different brittle mineral contents,this study pro-poses a method to identify the acoustic emission signal released by rock fracture under different brittle miner-al content(BMC),and then determine the content of brittle matter in rock.To understand related interference such as the noises in the acoustic emission signals released by the rock mass rupture,a 1DCNN-BLSTM network model with SE module is constructed in this study.The signal data is processed through the 1DCNN and BLSTM networks to fully extract the time-series correlation features of the signals,the non-correlated features of the local space and the weak periodicity law.Furthermore,the processed signals data is input into the fully connected layers.Finally,softmax function is used to accurately identify the acoustic emission signals released by different rocks,and then determine the content of brittle minerals contained in rocks.Through experimental comparison and analysis,1DCNN-BLSTM model embedded with SE module has good anti-noise performance,and the recognition accuracy can reach more than 90 percent,which is better than the traditional deep network models and provides a new way of thinking for rock acoustic emission re-search.
基金supported in part by the Major Program of the Ministry of Science and Technology of China under Grant 2019YFB2205102in part by the National Natural Science Foundation of China under Grant 61974164,62074166,61804181,62004219,62004220,62104256.
文摘With the continuous development of deep learning,Deep Convolutional Neural Network(DCNN)has attracted wide attention in the industry due to its high accuracy in image classification.Compared with other DCNN hard-ware deployment platforms,Field Programmable Gate Array(FPGA)has the advantages of being programmable,low power consumption,parallelism,and low cost.However,the enormous amount of calculation of DCNN and the limited logic capacity of FPGA restrict the energy efficiency of the DCNN accelerator.The traditional sequential sliding window method can improve the throughput of the DCNN accelerator by data multiplexing,but this method’s data multiplexing rate is low because it repeatedly reads the data between rows.This paper proposes a fast data readout strategy via the circular sliding window data reading method,it can improve the multiplexing rate of data between rows by optimizing the memory access order of input data.In addition,the multiplication bit width of the DCNN accelerator is much smaller than that of the Digital Signal Processing(DSP)on the FPGA,which means that there will be a waste of resources if a multiplication uses a single DSP.A multiplier sharing strategy is proposed,the multiplier of the accelerator is customized so that a single DSP block can complete multiple groups of 4,6,and 8-bit signed multiplication in parallel.Finally,based on two strategies of appeal,an FPGA optimized accelerator is proposed.The accelerator is customized by Verilog language and deployed on Xilinx VCU118.When the accelerator recognizes the CIRFAR-10 dataset,its energy efficiency is 39.98 GOPS/W,which provides 1.73×speedup energy efficiency over previous DCNN FPGA accelerators.When the accelerator recognizes the IMAGENET dataset,its energy efficiency is 41.12 GOPS/W,which shows 1.28×−3.14×energy efficiency compared with others.
