Providing autonomous systems with an effective quantity and quality of information from a desired task is challenging. In particular, autonomous vehicles, must have a reliable vision of their workspace to robustly acc...Providing autonomous systems with an effective quantity and quality of information from a desired task is challenging. In particular, autonomous vehicles, must have a reliable vision of their workspace to robustly accomplish driving functions. Speaking of machine vision, deep learning techniques, and specifically convolutional neural networks, have been proven to be the state of the art technology in the field. As these networks typically involve millions of parameters and elements, designing an optimal architecture for deep learning structures is a difficult task which is globally under investigation by researchers. This study experimentally evaluates the impact of three major architectural properties of convolutional networks, including the number of layers, filters, and filter size on their performance. In this study, several models with different properties are developed,equally trained, and then applied to an autonomous car in a realistic simulation environment. A new ensemble approach is also proposed to calculate and update weights for the models regarding their mean squared error values. Based on design properties,performance results are reported and compared for further investigations. Surprisingly, the number of filters itself does not largely affect the performance efficiency. As a result, proper allocation of filters with different kernel sizes through the layers introduces a considerable improvement in the performance.Achievements of this study will provide the researchers with a clear clue and direction in designing optimal network architectures for deep learning purposes.展开更多
One way to reduce energy consumption in wireless sensor networks is to reduce the number of active nodes in the network. When sensors are redundantly deployed, a subset of sensors should be selected to actively monito...One way to reduce energy consumption in wireless sensor networks is to reduce the number of active nodes in the network. When sensors are redundantly deployed, a subset of sensors should be selected to actively monitor the field (referred to as a "cover"), whereas the rest of the sensors should be put to sleep to conserve their batteries. In this paper, a learning automata based algorithm for energy-efficient monitoring in wireless sensor networks (EEMLA) is proposed. Each node in EEMLA algorithm is equipped with a learning automaton which decides for the node to be active or not at any time during the operation of the network. Using feedback received from neighboring nodes, each node gradually learns its proper state during the operation of the network. Experimental results have shown that the proposed monitoring algorithm in comparison to other existing methods such as Tian and LUC can better prolong the network lifetime.展开更多
多数基于深度学习的情感分类方法过于追求数据驱动,忽视了文本情感特征对于情感分类的影响;而一些融合情感特征的情感分类方法仅利用了部分相关特征,使得其他情感特征并没有得到充分利用.针对这一现象,提出了一种融合词级特征和句级特...多数基于深度学习的情感分类方法过于追求数据驱动,忽视了文本情感特征对于情感分类的影响;而一些融合情感特征的情感分类方法仅利用了部分相关特征,使得其他情感特征并没有得到充分利用.针对这一现象,提出了一种融合词级特征和句级特征的在线评论情感分类模型.首先利用构建的情感元素词典获取情感词、否定词等多种特征词,然后通过多维特征向量表示将多种文本特征转化为词级特征向量和句级特征向量,最后将这些特征向量融入LSTM网络模型完成情感分类,整个模型简称为MF-LSTM(sentiment classification model based on multidimensional features and LSTM).MF-LSTM充分利用了评论文本的情感先验知识,情感分类能力得到显著提升.在3个中文评论数据集上的实验结果表明MF-LSTM模型相比其他深度学习方法具有更好的分类效果,并且在样本数据不平衡的情况下也能具有较好的鲁棒性.展开更多
Text in natural scene images usually carries abundant semantic information. However, due to variations of text and complexity of background, detecting text in scene images becomes a critical and challenging task. In t...Text in natural scene images usually carries abundant semantic information. However, due to variations of text and complexity of background, detecting text in scene images becomes a critical and challenging task. In this paper, we present a novel method to detect text from scene images. Firstly, we decompose scene images into background and text components using morphological component analysis(MCA), which will reduce the adverse effects of complex backgrounds on the detection results.In order to improve the performance of image decomposition,two discriminative dictionaries of background and text are learned from the training samples. Moreover, Laplacian sparse regularization is introduced into our proposed dictionary learning method which improves discrimination of dictionary. Based on the text dictionary and the sparse-representation coefficients of text, we can construct the text component. After that, the text in the query image can be detected by applying certain heuristic rules. The results of experiments show the effectiveness of the proposed method.展开更多
The curse of dimensionality refers to the problem o increased sparsity and computational complexity when dealing with high-dimensional data.In recent years,the types and vari ables of industrial data have increased si...The curse of dimensionality refers to the problem o increased sparsity and computational complexity when dealing with high-dimensional data.In recent years,the types and vari ables of industrial data have increased significantly,making data driven models more challenging to develop.To address this prob lem,data augmentation technology has been introduced as an effective tool to solve the sparsity problem of high-dimensiona industrial data.This paper systematically explores and discusses the necessity,feasibility,and effectiveness of augmented indus trial data-driven modeling in the context of the curse of dimen sionality and virtual big data.Then,the process of data augmen tation modeling is analyzed,and the concept of data boosting augmentation is proposed.The data boosting augmentation involves designing the reliability weight and actual-virtual weigh functions,and developing a double weighted partial least squares model to optimize the three stages of data generation,data fusion and modeling.This approach significantly improves the inter pretability,effectiveness,and practicality of data augmentation in the industrial modeling.Finally,the proposed method is verified using practical examples of fault diagnosis systems and virtua measurement systems in the industry.The results demonstrate the effectiveness of the proposed approach in improving the accu racy and robustness of data-driven models,making them more suitable for real-world industrial applications.展开更多
传统机器学习和深度学习模型在处理情感分类任务时会忽略情感特征词的强度,情感语义关系单薄,造成情感分类的精准度不高。提出一种融合情感词典的改进型BiLSTM-CNN+Attention情感分类算法。首先,通过融合情感词典的特征提取方法优化特...传统机器学习和深度学习模型在处理情感分类任务时会忽略情感特征词的强度,情感语义关系单薄,造成情感分类的精准度不高。提出一种融合情感词典的改进型BiLSTM-CNN+Attention情感分类算法。首先,通过融合情感词典的特征提取方法优化特征词的权重;其次,利用卷积神经网络(convolutional neural network, CNN)提取局部特征,利用双向长短时记忆网络(bidirectional long and short-term memory, BiLSTM)高效提取上下文语义特征和长距离依赖关系;再结合注意力机制对情感特征加成;最后由Softmax分类器实现文本情感预测。实验结果表明:所提出的情感分类算法在精确率、召回率和F值上均有较大提升。相较于TextCNN、BiLSTM、长短时记忆网络(long and short-term memory, LSTM)、CNN和随机森林模型,所提方法的F值分别提高2.35%、3.63%、4.36%、2.72%和6.35%。这表明所提方法能够充分融合情感特征词的权重,利用上下文语义特征,提高情感分类性能。所提方法具有一定的学术价值和应用前景。展开更多
对注意力缺陷多动障碍症(ADHD)受试者的准确识别一直是神经科学研究和临床诊断的挑战。基于更好的区分正常人和患者这一目的,文中采用了一种基于字典学习和长短期记忆(Long Short Term Memory,LSTM)网络的Di-LSTM算法,通过利用快速独立...对注意力缺陷多动障碍症(ADHD)受试者的准确识别一直是神经科学研究和临床诊断的挑战。基于更好的区分正常人和患者这一目的,文中采用了一种基于字典学习和长短期记忆(Long Short Term Memory,LSTM)网络的Di-LSTM算法,通过利用快速独立成分分析(Fast Independent Component Analysis,FastICA)初始化的在线字典学习,获得相应时间序列并且结合LSTM进行分类实验,实验结果表明,所提方法分类准确率达到了79.01%,特异性为88.9%,灵敏度为62.7%,说明该方法对于识别ADHD患者有所帮助,具有较好的应用前景。展开更多
文摘Providing autonomous systems with an effective quantity and quality of information from a desired task is challenging. In particular, autonomous vehicles, must have a reliable vision of their workspace to robustly accomplish driving functions. Speaking of machine vision, deep learning techniques, and specifically convolutional neural networks, have been proven to be the state of the art technology in the field. As these networks typically involve millions of parameters and elements, designing an optimal architecture for deep learning structures is a difficult task which is globally under investigation by researchers. This study experimentally evaluates the impact of three major architectural properties of convolutional networks, including the number of layers, filters, and filter size on their performance. In this study, several models with different properties are developed,equally trained, and then applied to an autonomous car in a realistic simulation environment. A new ensemble approach is also proposed to calculate and update weights for the models regarding their mean squared error values. Based on design properties,performance results are reported and compared for further investigations. Surprisingly, the number of filters itself does not largely affect the performance efficiency. As a result, proper allocation of filters with different kernel sizes through the layers introduces a considerable improvement in the performance.Achievements of this study will provide the researchers with a clear clue and direction in designing optimal network architectures for deep learning purposes.
基金supported by the Islamic Azad University Urmia Brach,Iran
文摘One way to reduce energy consumption in wireless sensor networks is to reduce the number of active nodes in the network. When sensors are redundantly deployed, a subset of sensors should be selected to actively monitor the field (referred to as a "cover"), whereas the rest of the sensors should be put to sleep to conserve their batteries. In this paper, a learning automata based algorithm for energy-efficient monitoring in wireless sensor networks (EEMLA) is proposed. Each node in EEMLA algorithm is equipped with a learning automaton which decides for the node to be active or not at any time during the operation of the network. Using feedback received from neighboring nodes, each node gradually learns its proper state during the operation of the network. Experimental results have shown that the proposed monitoring algorithm in comparison to other existing methods such as Tian and LUC can better prolong the network lifetime.
文摘多数基于深度学习的情感分类方法过于追求数据驱动,忽视了文本情感特征对于情感分类的影响;而一些融合情感特征的情感分类方法仅利用了部分相关特征,使得其他情感特征并没有得到充分利用.针对这一现象,提出了一种融合词级特征和句级特征的在线评论情感分类模型.首先利用构建的情感元素词典获取情感词、否定词等多种特征词,然后通过多维特征向量表示将多种文本特征转化为词级特征向量和句级特征向量,最后将这些特征向量融入LSTM网络模型完成情感分类,整个模型简称为MF-LSTM(sentiment classification model based on multidimensional features and LSTM).MF-LSTM充分利用了评论文本的情感先验知识,情感分类能力得到显著提升.在3个中文评论数据集上的实验结果表明MF-LSTM模型相比其他深度学习方法具有更好的分类效果,并且在样本数据不平衡的情况下也能具有较好的鲁棒性.
基金supported in part by the National Natural Science Foundation of China(61302041,61363044,61562053,61540042)the Applied Basic Research Foundation of Yunnan Provincial Science and Technology Department(2013FD011,2016FD039)
文摘Text in natural scene images usually carries abundant semantic information. However, due to variations of text and complexity of background, detecting text in scene images becomes a critical and challenging task. In this paper, we present a novel method to detect text from scene images. Firstly, we decompose scene images into background and text components using morphological component analysis(MCA), which will reduce the adverse effects of complex backgrounds on the detection results.In order to improve the performance of image decomposition,two discriminative dictionaries of background and text are learned from the training samples. Moreover, Laplacian sparse regularization is introduced into our proposed dictionary learning method which improves discrimination of dictionary. Based on the text dictionary and the sparse-representation coefficients of text, we can construct the text component. After that, the text in the query image can be detected by applying certain heuristic rules. The results of experiments show the effectiveness of the proposed method.
基金supported in part by the National Natural Science Foundation of China(NSFC)(92167106,61833014)Key Research and Development Program of Zhejiang Province(2022C01206)。
文摘The curse of dimensionality refers to the problem o increased sparsity and computational complexity when dealing with high-dimensional data.In recent years,the types and vari ables of industrial data have increased significantly,making data driven models more challenging to develop.To address this prob lem,data augmentation technology has been introduced as an effective tool to solve the sparsity problem of high-dimensiona industrial data.This paper systematically explores and discusses the necessity,feasibility,and effectiveness of augmented indus trial data-driven modeling in the context of the curse of dimen sionality and virtual big data.Then,the process of data augmen tation modeling is analyzed,and the concept of data boosting augmentation is proposed.The data boosting augmentation involves designing the reliability weight and actual-virtual weigh functions,and developing a double weighted partial least squares model to optimize the three stages of data generation,data fusion and modeling.This approach significantly improves the inter pretability,effectiveness,and practicality of data augmentation in the industrial modeling.Finally,the proposed method is verified using practical examples of fault diagnosis systems and virtua measurement systems in the industry.The results demonstrate the effectiveness of the proposed approach in improving the accu racy and robustness of data-driven models,making them more suitable for real-world industrial applications.
文摘传统机器学习和深度学习模型在处理情感分类任务时会忽略情感特征词的强度,情感语义关系单薄,造成情感分类的精准度不高。提出一种融合情感词典的改进型BiLSTM-CNN+Attention情感分类算法。首先,通过融合情感词典的特征提取方法优化特征词的权重;其次,利用卷积神经网络(convolutional neural network, CNN)提取局部特征,利用双向长短时记忆网络(bidirectional long and short-term memory, BiLSTM)高效提取上下文语义特征和长距离依赖关系;再结合注意力机制对情感特征加成;最后由Softmax分类器实现文本情感预测。实验结果表明:所提出的情感分类算法在精确率、召回率和F值上均有较大提升。相较于TextCNN、BiLSTM、长短时记忆网络(long and short-term memory, LSTM)、CNN和随机森林模型,所提方法的F值分别提高2.35%、3.63%、4.36%、2.72%和6.35%。这表明所提方法能够充分融合情感特征词的权重,利用上下文语义特征,提高情感分类性能。所提方法具有一定的学术价值和应用前景。
文摘对注意力缺陷多动障碍症(ADHD)受试者的准确识别一直是神经科学研究和临床诊断的挑战。基于更好的区分正常人和患者这一目的,文中采用了一种基于字典学习和长短期记忆(Long Short Term Memory,LSTM)网络的Di-LSTM算法,通过利用快速独立成分分析(Fast Independent Component Analysis,FastICA)初始化的在线字典学习,获得相应时间序列并且结合LSTM进行分类实验,实验结果表明,所提方法分类准确率达到了79.01%,特异性为88.9%,灵敏度为62.7%,说明该方法对于识别ADHD患者有所帮助,具有较好的应用前景。