检索结果-内蒙古大学图书馆

automated intelligent system for sound signalling device quality assurance

INFORMATION SCIENCES 2015年 294卷 600-611页

作者： Maniak, Tomasz Jayne, Chrisina Iqbal, Rahat Doctor, Faiyaz UK NSI Co Ltd Nippon Seiki Engn & Qual Dept Redditch B98 9HL Worcs England Coventry Univ Fac Engn & Comp Coventry CV1 5FB W Midlands England

This paper presents a novel approach to the detection and recognition of faulty audio signalling devices as part of an automated industrial manufacturing quality assurance process. The proposed system outperforms other well-established automated systems based on mel-frequency cepstrum coefficients (MFCC) and multi-layer perceptron (MLP). It uses both unlabelled sound data and labelled historical data acquired from human experts in detecting faulty signalling devices. The unlabelled data is used to train a deep neural network generative model to create multiple levels of hierarchical feature extractors which are used to train an MLP classifier, with the intent to model the human reasoning and judging processes in respect to sound classification. This paper presents the results of real world experiments based on data pertaining to the audio signalling quality assurance process for car instrument cluster manufacturing. These results show that the proposed system is able to successfully classify speakers into two groups: "Good" and "No good" depending on the part quality. The proposed system proves to be capable enough to eliminate the need for a manual inspection within the manufacturing process and is shown to be able to diagnose a fault with a high degree of accuracy. This work can be extended to other areas of automotive inspection where there is a need for a robust solution to sound detection and where an output signal is represented by a complex and changing frequency spectrum even with significant environmental noise. (C) 2014 Elsevier Inc. All rights reserved.

关键词： Non-speech sound recognition Audio signal processing Artificial neural networks auto-encoder Manufacturing automation

来源：评论

学校读者我要写书评

暂无评论

Learning Feature Hierarchies: A Layer-Wise Tag-Embedded Approach

引用

IEEE TRANSACTIONS ON MULTIMEDIA 2015年第6期17卷 816-827页

作者： Yuan, Zhaoquan Xu, Changsheng Sang, Jitao Yan, Shuicheng Hossain, M. Shamim Chinese Acad Sci Inst Automat Natl Lab Pattern Recognit Beijing 100190 Peoples R China Natl Univ Singapore Dept Elect & Comp Engn Singapore 119077 Singapore King Saud Univ SWE Dept Coll Comp & Informat Sci Riyadh 11543 Saudi Arabia

Feature representation learning is an important and fundamental task in multimedia and pattern recognition research. In this paper, we propose a novel framework to explore the hierarchical structure inside the images from the perspective of feature representation learning, which is applied to hierarchical image annotation. Different from the current trend in multimedia analysis of using pre-defined features or focusing on the end-task "flat" representation, we propose a novel layer-wise tag-embedded deep learning (LTDL) model to learn hierarchical features which correspond to hierarchical semantic structures in the tag hierarchy. Unlike most existing deep learning models, LTDL utilizes both the visual content of the image and the hierarchical information of associated social tags. In the training stage, the two kinds of information are fused in a bottom-up way. Supervised training and multi-modal fusion alternate in a layer-wise way to learn feature hierarchies. To validate the effectiveness of LTDL, we conduct extensive experiments for hierarchical image annotation on a large-scale public dataset. Experimental results show that the proposed LTDL can learn representative features with improved performances.

关键词： auto-encoder deep learning hierarchical feature learning social tags

来源：评论

学校读者我要写书评

暂无评论

Study on Signal Recognition and Diagnosis for Spacecraft Based on Deep Learning Method

Study on Signal Recognition and Diagnosis for Spacecraft Bas...

引用

PROGNOSTICS AND SYSTEM HEALTH MANAGEMENT CONFERENCE (PHM)

作者： Li, Ke Wang, Quanxin Beijing Univ Aeronaut & Astronaut Sch Aeronaut Sci & Engn Beijing Peoples R China

ISBN: (纸本)9781467385541

According to the large variety of data generated during the spacecraft test and fault diagnosis, this paper designs a multi class classification algorithm based on deep learning method. The algorithm uses the stack auto-encoder to initialize the initial weights and offsets of the multi-layer neural network, and then monitor the parameters after the initialization with the gradient descent method. The algorithm can overcome many weaknesses of SVM, for example, it is too complex and occupied more space when data is large or the categories are huge. By studying the measured data, the expert knowledge can be provided for the spacecraft health management platform. Experimental data show that this depth learning algorithm can get a high accuracy in the classification of multi class signals of spacecraft.

关键词： Fault Diagnosis deep learning Data Compression Pattern Recognition auto-encoder

来源：评论

学校读者我要写书评

暂无评论

Person identification from streaming surveillance video using mid-level features from joint action-pose distribution

Person identification from streaming surveillance video usin...

引用

Conference on Video Surveillance and Transportation Imaging Applications

作者： Nair, Binu M. y Asari, Vijayan K. Univ Dayton ECE Dept Dayton OH 45469 USA

ISBN: (纸本)9781628414974

We propose a real time person identification algorithm for surveillance based scenarios from low-resolution streaming video, based on mid-level features extracted from the joint distribution of various types of human actions and human poses. The proposed algorithm uses the combination of an auto-encoder based action association framework which produces per-frame probability estimates of the action being performed, and a pose recognition framework which gives per-frame body part locations. The main focus in this manuscript is to effectively combine these per-frame action probability estimates and pose trajectories from a short temporal window to obtain mid-level features. We demonstrate that these mid-level features captures the variation in the action performed with respect to an individual and can be used to distinguish one person from the next. Preliminary analysis on the KTH action dataset where each sequence is annotated with a specific person and a specific action is provided and shows some interesting results which verify this concept.

关键词： auto-encoder Action Recognition Restricted Boltzmann Machines Histogram of Flow Local Binary Flow Patterns Pose Trajectories

来源：评论

学校读者我要写书评

暂无评论

DEEP SEMI-SUPERVISED LEARNING FOR DOMAIN ADAPTATION 25

DEEP SEMI-SUPERVISED LEARNING FOR DOMAIN ADAPTATION

引用

IEEE International Workshop on Machine Learning for Signal Processing

作者： Chen, Hung-Yu Chien, Jen-Tzung Natl Chiao Tung Univ Dept Elect & Comp Engn Hsinchu 30010 Taiwan

ISBN: (纸本)9781467374545

Domain adaptation aims to adapt a classifier from source domain to target domain through learning a good feature representation that allows knowledge to be shared and transferred across domains. Most of previous studies are restricted to extract features and train classifier separately under a shallow model structure. In this paper, we propose a semi-supervised domain adaptation method which co-trains the feature representation and pattern classification under deep neural network (DNN) framework. The labeling in target domain is not required. We treat the hidden layers in DNN as feature extraction and construct the output layer consisting of classification and regression. Our idea is to conduct the feature-based domain adaptation which jointly minimizes the divergence between the distributions from labeled and unlabeled data in both domains, the reconstruction errors due to an autoencoder, and the classification errors due to the labeled data in source domain. Experiments on image recognition and sentiment classification show the superiority of DNN co-training for domain adaptation.

关键词： Deep learning domain adaptation semi-supervised learning co-training auto-encoder

来源：评论

学校读者我要写书评

暂无评论

Learning deep dynamical models from image pixels

Learning deep dynamical models from image pixels

引用

17th IFAC Symposium on System Identification (SYSID 2015)

作者： Wahlstrom, Niklas Schon, Thomas B. Deisenroth, Marc Peter Linkoping Univ Dept Elect Engn S-58183 Linkoping Sweden Uppsala Univ Dept Informat Technol Uppsala Sweden Univ London Imperial Coll Sci Technol & Med Dept Comp London SW7 2AZ England

Modeling dynamical systems is important in many disciplines, such as control, robotics, or neurotechnology. Commonly the state of these systems is not directly observed, but only available through noisy and potentially high-dimensional observations. In these cases, system identification, i.e., finding the measurement mapping and the transition mapping (system dynamics) in latent, space can be challenging. For linear system dynamics and measurement mappings efficient solutions for system identification are available. However, in practical applications, the linearity assumptions does not hold, requiring nonlinear system identification techniques. If additionally the observations are high-dimensional (e.g., images), nonlinear system identification is inherently hard. To address the problem of nonlinear system identification from high-dimensional observations, we combine recent advances in deep learning and system identification. In particular, we jointly learn a low-dimensional embedding of the observation by means of deep auto-encoders and a predictive transition model in this low-dimensional space. We demonstrate that our model enables learning good predictive models of dynamical systems from pixel information only. (C) 2015, IFAC (International Federation of automatic Control) Hosting by Elsevier Ltd. All rights reserved.

关键词： Deep neural networks system identification nonlinear systems low-dimensional embedding auto-encoder

来源：评论

学校读者我要写书评

暂无评论

automatic Document Summarization via Deep Neural Networks 8

Automatic Document Summarization via Deep Neural Networks

引用

8th International Symposium on Computational Intelligence and Design (ISCID)

作者： Yao, Chengwei Shen, Jianfen Chen, Gencai Zhejiang Univ Coll Comp Sci & Technol Hangzhou Zhejiang Peoples R China Hlth Informat Ctr Zhejiang Hangzhou Zhejiang Peoples R China

ISBN: (纸本)9781467395878

automatic document summarization aim to extracting sentences which might cover the main content of a document or documents. To achieve this, many algorithms have been tried to rank the sentences by using task-specific features in a shallow architecture. The main challenge of these approaches is to keep balance between information coverage and redundancy because of absence of discovering the intrinsic semantic representation. Inspired by the recent successful achievement of Deep Learning, this paper proposes a new framework of document summarization via Deep Neural Networks ( DNNs). Specifically, we feed the sentences as the input to the visible layer of DNNs. After pretraining layer by layer and fine-tuning, the lower dimensional semantic space can be revealed. Based on this space, we design sentences extraction algorithm to construct the summary. Experiments on the DUC2006 and DUC2007 dataset show that our framework works better than state-ofthe-art methods.

关键词： Document Summarization Deep Learning Deep Brief Network Restricted Boltzmann Machine auto-encoder ROUGE

来源：评论

学校读者我要写书评

暂无评论

Novel Hierarchical Cellular Simultaneous Recurrent Neural Network for Object Detection

Novel Hierarchical Cellular Simultaneous Recurrent Neural Ne...

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Alam, M. Vidyaratne, L. Iftekharuddin, K. M. Old Dominion Univ Dept Elect & Comp Engn Vis Lab Norfolk VA 23529 USA

ISBN: (纸本)9781479919598

Large scale feed forward neural networks have seen intense application in many computer vision problems. However, these networks can get hefty and computationally intensive with increasing complexity of the task. This work, for the first time in literature, introduces a Cellular Simultaneous Recurrent Network (CSRN) based hierarchical neural network for object detection. CSRN has shown to be more effective to solving complex tasks such as maze traversal and image processing when compared to generic feed forward networks. While deep neural networks (DNN) have exhibited excellent performance in object detection and recognition, such hierarchical structure has largely been absent in neural networks with recurrency. This work attempts to introduce deep hierarchy in CSRN for object detection. We propose a novel CSRN based DNN feature extractor that utilizes highly efficient filters derived from CSRNs used in the hidden convolutional layer. Experiments using a face detection task show that the CSRN based DNN offers comparable performance to the state-of-the-art convolutional neural network (CNN), while utilizing only as few as five filters in each of its hidden layers. In comparison, the CNN requires from a few to thousands of filters in each of its hidden layers for the same task. We also demonstrate the flexibility of the proposed architecture by introducing hybridization concept to the network to enhance its scale invariance. The hybrid scale invariant architecture is tested with randomly scaled object image dataset for face detection. Finally, we show that the CSRN based hybrid DNN performance is also comparable to that of the hybrid CNN.

关键词： Cellular Simultaneous Recurrent Network (CSRN) Hierarchical Recurrent Neural Network Unsupervised Learning Feature extraction auto-encoder

来源：评论

学校读者我要写书评

暂无评论

Recognizing Objectionable Images Using Convolutional Neural Nets

Recognizing Objectionable Images Using Convolutional Neural ...

引用

Signal Processing and Intelligent Systems Conference (SPIS)

作者： Moradi, Reza Yousefzadeh, Rahman Iran Univ Sci & Technol Tehran Iran Shahrood Univ Technol Shahrood Iran

ISBN: (纸本)9781509001392

In recent years different methods for detecting objectionable images have proposed. All of the previous systems are based on extracting pre-defined and certain features from the images. In this paper a method is proposed in order to detect objectionable images using convolutional neural networks. In this method first features are learned through a sparse auto-encoder and then training is done by a convolutional neural network. The architecture of the network consists of convolution and sub-sampling layers followed by a fully connected output layer which feeds a softmax classifier with cross entropy cost function. The proposed method is able to effectively detect 90.5% of images correctly employing a rather small training dataset.

关键词： deep learning convolutional neural network objectionable image recognition auto-encoder

来源：评论

学校读者我要写书评

暂无评论

Learning Agent for a Heat-Pump Thermostat with a Set-Back Strategy Using Model-Free Reinforcement Learning

引用

ENERGIES 2015年第8期8卷 8300-8318页

作者： Ruelens, Frederik Iacovella, Sandro Claessens, Bert J. Belmans, Ronnie Katholieke Univ Leuven Dept Elect Engn Fac Engn Div ELECTA Kasteelpk Arenberg 10Box 2445 B-3001 Leuven Belgium EnergyVille B-3600 Genk Belgium Flemish Inst Technol Res Energy Dept VITO B-2400 Mol Belgium

The conventional control paradigm for a heat pump with a less efficient auxiliary heating element is to keep its temperature set point constant during the day. This constant temperature set point ensures that the heat pump operates in its more efficient heat-pump mode and minimizes the risk of activating the less efficient auxiliary heating element. As an alternative to a constant set-point strategy, this paper proposes a learning agent for a thermostat with a set-back strategy. This set-back strategy relaxes the set-point temperature during convenient moments, e.g., when the occupants are not at home. Finding an optimal set-back strategy requires solving a sequential decision-making process under uncertainty, which presents two challenges. The first challenge is that for most residential buildings, a description of the thermal characteristics of the building is unavailable and challenging to obtain. The second challenge is that the relevant information on the state, i.e., the building envelope, cannot be measured by the learning agent. In order to overcome these two challenges, our paper proposes an auto-encoder coupled with a batch reinforcement learning technique. The proposed approach is validated for two building types with different thermal characteristics for heating in the winter and cooling in the summer. The simulation results indicate that the proposed learning agent can reduce the energy consumption by 4%-9% during 100 winter days and by 9%-11% during 80 summer days compared to the conventional constant set-point strategy.

关键词： auto-encoder batch reinforcement learning heat pump set-back thermostat

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：