检索结果-内蒙古大学图书馆

Language/Dialect Recognition Based on Unsupervised Deep Learning

IEEE-ACM TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING 2018年第5期26卷 873-882页

作者： Zhang, Qian Hansen, John H. L. Univ Texas Dallas Erik Jonsson Sch Engn Ctr Robust Speech Syst Richardson TX 75080 USA

Over the past decade, bottleneck features within an i-Vector framework have been used for state-of-the-art language/dialect identification (LID/DID). However, traditional bottleneck feature extraction requires additional transcribed speech information. Alternatively, two types of unsupervised deep learning methods are introduced in this study. To address this limitation, an unsupervised bottleneck feature extraction approach is proposed, which is derived from the traditional bottleneck structure but trained with estimated phonetic labels. In addition, based on a generative modeling autoencoder, two types of latent variable learning algorithms are introduced for speech feature processing, which have been previous considered for image processing/reconstruction. Specifically, a variational autoencoder and adversarial autoencoder are utilized on alternative phase of speech processing. To demonstrate the effectiveness of the proposed methods, three corpora are evaluated: 1) a four Chinese dialect dataset, 2) a five Arabic dialect corpus, and 3) multigenre broadcast challenge corpus (MGB-3) for arabic DID. The proposed features are shown to outperform traditional acoustic feature MFCCs consistently across three corpora. Taken collectively, the proposed features achieve up to a relative +58% improvement in Cavg for LID/DID without the need of any secondary speech corpora.

关键词： Language/Dialect recognition unsupervised learning variational autoencoder adversarial autoencoder bottleneck feature phonetic label estimation

来源：评论

学校读者我要写书评

暂无评论

Multi-view face generation via unpaired images

引用

VISUAL COMPUTER 2022年第7期38卷 2539-2554页

作者： Wang, Shuai Zou, Yanni Min, Weidong Wu, Jiansheng Xiong, Xin Nanchang Univ Sch Informat Engn Nanchang 330031 Jiangxi Peoples R China Nanchang Univ Sch Software Nanchang 330047 Jiangxi Peoples R China Jiangxi Key Lab Smart City Nanchang 330047 Jiangxi Peoples R China

Multi-view face generation from a single image is an essential and challenging problem. Most of the existing methods need to use paired images when training models. However, collecting and labeling large-scale paired face images could lead to high labor and time cost. In order to address this problem, multi-view face generation via unpaired images is proposed in this paper. To avoid using paired data, the encoder and discriminator are trained, so that the high-level abstract features of the identity and view of the input image are learned by the encoder, and then, these low-dimensional data are input into the generator, so that the realistic face image can be reconstructed by the training generator and discriminator. During testing, multiple one-hot vectors representing the view are imposed to the identity representation and the generator is employed to map them to high-dimensional data, respectively, which can generate multi-view images while preserving the identity features. Furthermore, to reduce the number of used labels, semi-supervised learning is used in the model. The experimental results show that our method can produce photo-realistic multi-view face images with a small number of view labels, and makes a useful exploration for the synthesis of face images via unpaired data and very few labels.

关键词： Multi-view face generation Generative adversarial net adversarial autoencoder Semi-supervised learning

来源：评论

学校读者我要写书评

暂无评论

An adversarial Neuro-Tensorial Approach for Learning Disentangled Representations

引用

INTERNATIONAL JOURNAL OF COMPUTER VISION 2019年第6-7期127卷 743-762页

作者： Wang, Mengjiao Shu, Zhixin Cheng, Shiyang Panagakis, Yannis Samaras, Dimitris Zafeiriou, Stefanos Imperial Coll London London England SUNY Stony Brook Stony Brook NY 11794 USA Middlesex Univ London England

Several factors contribute to the appearance of an object in a visual scene, including pose, illumination, and deformation, among others. Each factor accounts for a source of variability in the data, while the multiplicative interactions of these factors emulate the entangled variability, giving rise to the rich structure of visual object appearance. Disentangling such unobserved factors from visual data is a challenging task, especially when the data have been captured in uncontrolled recording conditions (also referred to as in-the-wild) and label information is not available. In this paper, we propose a pseudo-supervised deep learning method for disentangling multiple latent factors of variation in face images captured in-the-wild. To this end, we propose a deep latent variable model, where the multiplicative interactions of multiple latent factors of variation are explicitly modelled by means of multilinear (tensor) structure. We demonstrate that the proposed approach indeed learns disentangled representations of facial expressions and pose, which can be used in various applications, including face editing, as well as 3D face reconstruction and classification of facial expression, identity and pose.

关键词： adversarial autoencoder Disentangled representation Tensor decomposition

来源：评论

学校读者我要写书评

暂无评论

TFAM-AAE-Uk: A Dual-Metric Spectrum Anomaly Detection Algorithm

引用

IEEE COMMUNICATIONS LETTERS 2024年第11期28卷 2638-2642页

作者： Ji, Haipeng Zhang, Tao Qiao, Xiaoqiang Wu, Hao Gui, Guan Nanjing Univ Informat Sci & Technol Sch Elect & Informat Engn Nanjing 210044 Peoples R China Natl Univ Def Technol Res Inst 63 Nanjing 210007 Peoples R China Nanjing Univ Posts & Telecommun Coll Telecommun & Informat Engn Nanjing 210003 Peoples R China

Effective spectrum management critically depends on the ability to detect anomalies caused by both legal user (LU) violations and illegal user (IU) intrusions. In this study, we introduce TFAM-AAE-U-k, an innovative spectrum anomaly detection framework that integrates an adversarial autoencoder (AAE) with a Time Frequency Attention Mechanism (TFAM) and a specialized user discriminator, U-k. This framework first employs the TFAM-AAE to reconstruct spectrum data by exploiting latent, time, and frequency features of the input. The reconstruction error serves as the primary anomaly detection metric. To address the challenges posed by low interference-to- signal ratio (ISR) IUs and camouflaged LUs, U-k is utilized to extract fingerprint features vectors from IQ data, establishing class center vectors for each LU. The Mahalanobis distance between these vectors is then calculated as a secondary metric for anomaly detection. Our experimental results demonstrate that TFAM-AAE-U-k consistently achieves excellent detection performance, particularly in scenarios involving challenging and hard-to-detect anomalies.

关键词： Feature extraction Vectors Anomaly detection Time-frequency analysis Training Fingerprint recognition Law Spectrum anomaly detection adversarial autoencoder time frequency attention mechanism openset recognition

来源：评论

学校读者我要写书评

暂无评论

EAAE: A Generative adversarial Mechanism Based Classfication Method for Small-scale Datasets

引用

NEURAL PROCESSING LETTERS 2023年第2期55卷 969-987页

作者： Chen, Ping Deng, Yuhui Zou, Qiang Lu, Lijuan Li, Hong Jinan Univ Dept Comp Sci Guangzhou 510632 Peoples R China Southwest Univ Dept Comp Sci Chongqing 400715 Peoples R China South China Univ Technol Sch Business Adm Guangzhou 510640 Peoples R China

When used for small-scale datasets classification tasks, deep neural networks are difficult to train, which results in the network not extracting useful features and the low accuracy of network. This paper proposes a method called Enhanced adversarial autoencoders (EAAE) to improve the accuracy of classification in small-scale datasets scenarios. We reuse the decoder of the adversarial autoencoders to generate data for the specified category, which expands the training space of the samples. In order to capture the features of the generated data, we also reuse the encoder to generate the corresponding latent label variables and train them in a supervised manner. To maintain and improve the generative power of the decoder, we impose conditional adversarial training on the generated data. Finally, we use the encoder with latent feature extraction capability to complete the classification by transfer learning. To evaluate the performance of the proposed method, we validate on a real Gastric Cancer dataset and five widely used UCI public datasets. Extensive experiments show that EAAE improves significantly in classification accuracy on small-scale datasets and performs better on large-scale datasets compared to popular classification methods. Furthermore, we evaluate the time cost of EAAE and conduct ablation experiments to demonstrate the effectiveness of the various modules of EAAE.

关键词： Small-scale datasets Deep learning adversarial autoencoder Generative adversarial network Classification tasks

来源：评论

学校读者我要写书评

暂无评论

Incorporating Geological Knowledge into Deep Learning to Enhance Geochemical Anomaly Identification Related to Mineralization and Interpretability

引用

MATHEMATICAL GEOSCIENCES 2024年第6期56卷 1233-1254页

作者： Zhang, Chunjie Zuo, Renguang China Univ Geosci State Key Lab Geol Proc & Mineral Resources Wuhan 430074 Peoples R China

Effective geochemical anomaly identification is crucial in mineral exploration. Recent trends have favored deep learning (DL) to decipher geochemical survey data. Yet purely data-driven DL algorithms often lack logical explanations and geological consistency, occasionally clashing with known geological insights and complicating model interpretation. A deep understanding of the geological processes forming the target mineral deposit is vital for accurate anomaly detection. Here, we introduce an adversarial autoencoder (AAE) network that integrates prior geological knowledge to identify geochemical anomalies linked to tungsten mineralization in southern Jiangxi Province, China. Considering the geochemical patterns linked to tungsten mineralization, Yanshanian granites and faults were strategically chosen as ore-controlling factors. The methodology employed multifractal singularity analysis to quantitatively measure the correlations between these ore-controlling factors and known tungsten deposits, aiming to establish an ore-forming regularity. This regularity serves as a priori distribution to control the encoder network's latent vector, refining the model's output. A comparison of detected geochemical anomalies under different constraints (AAE, Granite_AAE, Fault_AAE, and Fault_Granite_AAE) revealed that AAE models incorporating prior geological information consistently outperformed unconstrained models in terms of anomaly detection. Integrating geological expertise with DL, our study overcomes the challenges of models relying purely on data or theory, offering a promising approach to geochemical exploration.

关键词： adversarial autoencoder Latent variables Geochemical anomalies detection Prior geological knowledge

来源：评论

学校读者我要写书评

暂无评论

Hyperspectral target detection using self-supervised background learning

引用

ADVANCES IN SPACE RESEARCH 2024年第2期74卷 628-646页

作者： Ali, Muhammad Khizer Amin, Benish Maud, Abdur Rahman Bhatti, Farrukh Aziz Sukhia, Komal Nain Khurshid, Khurram Inst Space Technol Dept Elect Engn iVis Lab Islamabad 44000 Pakistan Bahria Univ Ctr Excellence Artificial Intelligence CoE AI Islamabad 44000 Pakistan Alf Ain Technol Pvt Ltd Lahore 54890 Pakistan NewVat Technol Pvt Ltd Islamabad 44000 Pakistan

Hyperspectral target detection is challenging in scenarios where spectral variability is high due to noise, spectral redundancy, and mixing. In addition, this spectral variability also creates the need for target detection algorithms to be robust against variations in the detection threshold. To overcome these challenges, this paper proposes a novel two-stage process for improved target detection in hyperspectral data. In the first stage, coarse detection is performed using a detector with a high probability of detection to identify background samples. These background samples are then used for background learning using an adversarial autoencoder (AAE) network, having spectral angle mapper (SAM) and Huber loss functions to minimize the impact of target pixels' contamination. In the second stage, an inference is made using the spectral difference between the hyperspectral data and the output of the learned background model, which helps in reducing the false alarm rate of the first stage. The proposed approach is compared with seven other target detection techniques using multiple datasets and evaluated through several metrics, such as the area under the curve (AUC) and signal-to-noise probability ratio (SNPR). Results reveal that the proposed technique outperforms other detectors in terms of SNPR, indicating improved target detectability, background suppressibility, and more tolerance to variations in the detection threshold. (c) 2024 COSPAR. Published by Elsevier B.V. All rights reserved.

关键词： Remote sensing Hyperspectral target detection Self -supervised learning adversarial autoencoder

来源：评论

学校读者我要写书评

暂无评论

Visual Interpretable Deep Learning Algorithm for Geochemical Anomaly Recognition

引用

NATURAL RESOURCES RESEARCH 2022年第5期31卷 2211-2223页

作者： Luo, Zijing Zuo, Renguang Xiong, Yihui China Univ Geosci State Key Lab Geol Proc & Mineral Resources Wuhan 430074 Peoples R China

Deep learning algorithms (DLAs) have achieved better results than traditional methods in the field of multivariate geochemical anomaly recognition because of their strong ability to extract feature from nonlinear data. However, most of DLAs are black-box approaches because of the high nonlinearity characteristics of the hidden layer. In addition, the integration of domain knowledge into the DLAs to ensure physical consistency is a challenge for DLAs in geoscience. In this study, we adopted the adversarial autoencoder (AAE) algorithm for geochemical anomaly detection. The interpretability of the model is improved by visualizing features and integrating geological domain knowledge into the loss function of the AAE. The feature visualization method was used to display the changes of information in the model calculation process to further understand the inherent operation law and principle of the neural network. The penalty term was added to the optimized loss function, and the spatiotemporal and genetic relationships between felsic intrusions and mineralization were integrated into the AAE with the aim of improving the geological interpretability of the network. The added penalty item can guide the changes in the stage of data reconstruction and improve the understandability of the results of geologically constrained AAE. In addition, the effectiveness of injecting the concept of physical constraints into the AAE can be verified via feature visualization. A case study in the southern Jiangxi Province and its surrounding areas was performed to identify multivariate geochemical anomalies. The results obtained by the geologically constrained AAE demonstrated a strong spatial correlation with the outcrop of intrusions in the study area, and most of the known mineral deposits are located in or near the highly anomalous areas.

关键词： Geochemical exploration Deep learning adversarial autoencoder Feature visualization Physics-constrained deep learning

来源：评论

学校读者我要写书评

暂无评论

Self-Supervised Learning Based Anomaly Detection in Synthetic Aperture Radar Imaging

IEEE OPEN JOURNAL OF SIGNAL PROCESSING

引用

IEEE OPEN JOURNAL OF SIGNAL PROCESSING 2022年 3卷 440-449页

作者： Muzeau, Max Ren, Chengfang Angelliaume, Sebastien Datcu, Mihai Ovarlez, Jean-Philippe Univ Paris Saclay SONDRA CentraleSupelec F-91190 Gif Sur Yvette France Univ Paris Saclay DEMR CentraleSupelec F-91190 Gif Sur Yvette France Univ Politehn Bucharest UPB Romania & German Aerosp Ctr DLR D-82234 Wessling Germany

In this paper, we proposed to investigate unsupervised anomaly detection in Synthetic Aperture Radar (SAR) images. Our approach considers anomalies as abnormal patterns that deviate from their surroundings without prior knowledge of their characteristics. This method deals with the crucial problems related to the presence of speckle, the spatial correlation structures in SAR images, and the lack of annotated data to train a detection algorithm. Our proposed method aims to address these issues through a self-supervised learning algorithm. First, we propose to mitigate the SAR speckle through the deep learning SAR2SAR algorithm. We then develop an adversarial autoencoder (AAE) to reconstruct anomaly-free SAR images from despeckled data taking into account potential spatial correlation structures. Finally, a change detection processing step is applied between the input and the output to detect anomalies. Experiments are performed to show the advantages of our method compared to the conventional Reed-Xiaoli algorithm, highlighting the importance of an efficient despeckling pre-processing step.

关键词： Radar polarimetry Speckle Synthetic aperture radar Anomaly detection Correlation Spaceborne radar Signal processing algorithms adversarial autoencoder anomaly detection deep-learning despeckling SAR self-supervised

来源：评论

学校读者我要写书评

暂无评论

Root cause analysis of manufacturing variation from optical scanning data

引用

ANNALS OF OPERATIONS RESEARCH 2024年第1-2期339卷 111-130页

作者： Bui, Anh Tuan Virginia Commonwealth Univ Dept Stat Sci & Operat Res Richmond VA 23284 USA

Identifying the root causes of part-to-part variation is a central problem in most six-sigma programs, especially of modern manufacturing processes. This is challenging as the sources and patterns of the variation are often unknown or previously unidentified. A small literature aims to address this problem by discovering unknown, previously unidentified variation sources, in a manner that helps understand their nature, from only a sample of measurement data. However, the common solution of this literature is unideal for this objective in terms of both methodology and metrology aspects. This paper proposes a convolutional generative modeling framework for optical scanning data to address this limitation. The proposed approach can correctly discover the true variation sources and visualize their individual patterns in two manufacturing examples, without any prior knowledge of the variation. The approach also outperforms state-of-the-art methods in these examples.

关键词： adversarial autoencoder Convolutional neural networks Generative adversarial networks (GAN) Phase I analysis Statistical process control Variation reduction

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：