检索结果-内蒙古大学图书馆

作者： Bao, Ruying Princeton University

学位级别：Ph.D., Doctor of Philosophy

Deep neural networks are widely used in signal processing from a broad range of areas due to their good performances, including computer vision, natural language processing, automatic driving, and so on. However, people notice that neural networks are easily fooled by adversarial attacks and very sensitive to certain data-related scenarios, such as imbalanced classes and outliers. In this thesis, we focus on enhancing model robustness of deep neural networks from different data distributions. In the first part, we focus on datasets whose distributions are biased naturally, from data collection or the nature of data. We define novel information-entropy-based classification loss functions (entropy weight and entropy noise) to distinguish the difficulty of each sample prediction by either weighting or introducing stochastic noise on top of the cross entropy loss. To evaluate the effectiveness of each loss function, we test the new loss functions on crafted noisy and imbalanced datasets based on MNIST. To illustrate their effectiveness in real scenarios, we show improvements on tasks including computer vision and natural language understanding, compared to the corresponding state of the art (SOTA) models. The results show that models trained with entropy-based loss functions surpass the SOTA models. Deep neural networks have also been demonstrated to be vulnerable to adversarial attacks, where small perturbations intentionally added to the original inputs can fool the classifier. In the second part, we propose Path-Norm regularization to improve robustness of neural networks against adversarial attacks in various Lp norms. By adding Path-Norm regularization, models achieve comparable performance as the SOTA defense methods, and outperform SOTA methods when attacks and training samples are from different Lp spaces. We also introduce Featurized Bidirectional Generative Adversarial Networks (FBGAN), which extracts semantic features of inputs and filters the non-semantic pe

关键词： Adversarial defense methods Deep learning Entropy-based objective functions Generative model Reconstructive defense methods Regularizing neural networks

来源：评论

学校读者我要写书评

暂无评论

Deep neural Decision Forest for Acoustic Scene Classification

Deep Neural Decision Forest for Acoustic Scene Classificatio...

引用

European signal processing Conference (EUSIPCO)

作者： Jianyuan Sun Xubo Liu Xinhao Mei Jinzheng Zhao Mark D. Plumbley Volkan Kılıç Wenwu Wang Centre for Vision Speech and Signal Processing (CVSSP) University of Surrey UK College of Computer Science and Technology Qingdao University China Department of Electrical and Electronics Engineering Izmir Katip Celebi University Turkey

ISBN: (数字)9789082797091

ISBN: (纸本)9781665467995

Acoustic scene classification (ASC) aims to classify an audio clip based on the characteristic of the recording environment. In this regard, deep learning based approaches have emerged as a useful tool for ASC problems. Conventional approaches to improving the classification accuracy include integrating auxiliary methods such as attention mechanism, pre-trained models and ensemble multiple sub-networks. However, due to the complexity of audio clips captured from different environments, it is difficult to distinguish their categories without using any auxiliary methods for existing deep learning models using only a single classifier. In this paper, we propose a novel approach for ASC using deep neural decision forest (DNDF). DNDF combines a fixed number of convolutional layers and a decision forest as the final classifier. The decision forest consists of a fixed number of decision tree classifiers, which have been shown to offer better classification performance than a single classifier in some datasets. In particular, the decision forest differs substantially from traditional random forests as it is stochastic, differentiable, and capable of using the back-propagation to update and learn feature representations in neural network. Experimental results on the DCASE2019 and ESC-50 datasets demonstrate that our proposed DNDF method improves the ASC performance in terms of classification accuracy and shows competitive performance as compared with state-of-the-art baselines.

关键词： Deep learning image analysis neural networks stochastic processes Training data Forestry Acoustics

来源：评论

学校读者我要写书评

暂无评论

Enhancing Arabic Text Classification: A Comparative Study of Machine Learning and Deep Learning Approaches

Enhancing Arabic Text Classification: A Comparative Study of...

引用

International Symposium on signal, image, Video and Communications (ISIVC)

作者： Idriss Akhadam Habib Ayyad Mathematics Computer Science & Applications Laboratory Hassan 2 university Mohammedia Morocco

ISBN: (数字)9798350385267

ISBN: (纸本)9798350385274

This paper presents a comprehensive pipeline that integrates machine learning (ML) and deep learning (DL) methods for Arabic text classification, achieving high accuracy. Our approach includes a detailed preprocessing phase,that involves tokenization, removal of non-Arabic words, stop words, punctuation, normalization of the Hamza character, and an optional stemming process. We utilize Bag of Words (BOW) and Term Frequency-Inverse Document Frequency (TF-IDF) for feature selection. Various ML and DL algorithms have been explored using the AL-Khaleej database, which contains 45500 arabic news documents. Performance evaluation on the testing set focuses on accuracy, precision, recall, and F1-score metrics. Our findings reveal that Logistic Regression and stochastic Gradient Descent (SGD) classifiers perform robustly with both feature extraction methods, achieving 97% accuracy. Linear Support Vector Classifier (SVC) shows a slight improvement with TF-IDF. Stemming’s impact varies, with SVC improving to 98% accuracy without stemming using BOW. The Convolutional neural Network (CNN) models, particularly with word-level representation, achieve superior accuracy of 97%, indicating their effectiveness in capturing contextual nuances and semantic relationships. Our pipeline not only accurately classifies Arabic texts but also provides insights into the efficacy of preprocessing and feature selection techniques in Arabic language processing. The success of our pipeline underscores its potential as a pivotal contribution to the advancement of Arabic text classification.

关键词： Deep learning Accuracy Text categorization Semantics Pipelines Static VAr compensators Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Radio-astronomical image Reconstruction with Conditional Denoising Diffusion Model

arXiv

引用

arXiv 2024年

作者： Drozdova, Mariia Kinakh, Vitaliy Bait, Omkar Taran, Olga Lastufka, Erica Dessauges-Zavadsky, Miroslava Holotyak, Taras Schaerer, Daniel Voloshynovskiy, Slava Department of Computer Science University of Geneva 7 route de Drize Carouge1227 Switzerland Geneva Observatory University of Geneva 51 Chemin Pegasi Versoix1290 Switzerland

Reconstructing sky models from dirty radio images for accurate source extraction, e.g. source localisation and flux estimation, is a complex yet critical task. It has important applications in galaxy evolution studies at high redshift, particularly in deep extragalactic fields using for example the Atacama Large Millimetre Array (ALMA). With the development of large-scale projects like the Square Kilometre Array (SKA), we anticipate the need for more advanced source extraction methods. Existing techniques, such as CLEAN and PyBDSF, currently struggle to effectively extract faint sources, highlighting the necessity for the development of more precise and robust methods. The success of source extraction task critically depends on the quality and accuracy of image reconstruction. Since the imaging process represents an information-lossy operator, the reconstruction is characterized by uncertainty. The current study proposes the application of stochastic neural networks for the direct reconstruction of sky models from dirty images. This approach allows to localize radio sources and to determine their fluxes with the corresponding uncertainties, providing a potential advancement in the field of radio source characterization. We utilize a dataset of 10164 images simulated with the CASA tool simalma based on the ALMA Cycle 5.3 antenna configuration. We apply conditional Denoising Diffusion Probabilistic Models (DDPMs) to directly reconstruct sky models from these dirty images, followed by processing through Photutils to extract the coordinates and fluxes of the sources. To test the robustness of the proposed model which was trained on a fixed water vapour value, we examine its performance under varying levels of water vapor. The proposed approach demonstrates state-of-the-art in source localisation, achieving over 90% completeness at a signal-to-noise ratio (SNR) as low as 2. Additionally, the described method offers an inherent measure of prediction reliability thanks to

关键词： image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Deep emotion recognition based on audio-visual correlation

引用

IET computer vision 2020年第7期14卷 517-527页

作者： Hajarolasvadi, Noushin Demirel, Hasan Eastern Mediterranean Univ Dept Elect & Elect Engn 10 Via Mersin TR-99628 Cyprus Turkey

Human emotion recognition is studied by means of unimodal channels over the last decade. However, efforts continue to answer tempting questions about how variant modalities can complement each other. This study proposes a multimodal approach using three-dimensional (3D) convolutional neural networks (CNNs) to model human emotion through a modality-referenced system while investigating the solution to such questions. The proposed modality-referenced system selects the input data based on one of the modalities regarded as reference or master. The other modality which is referred to as a slave simply adjusts or attunes itself with the master in the temporal domain. In this context, the authors developed three multimodal emotion recognition system, namely, video-referenced system, audio-referenced system, and the audio-visual-referenced system to explore the congruence impact of audio and video modalities on each other. Two pipelines of 3D CNN architectures are employed where k-means clustering is used in the master pipeline and the slave pipeline adapts itself in a temporal sense. The outputs of the two pipelines are fused to improve recognition performance. In addition, canonical correlation analysis and t-distributed stochastic neighbour embedding is used validating the experiments. Results show that temporal alignment of the data between two modalities improves the recognition performance significantly.

关键词： deep emotion recognition video modalities audio–visual-referenced system slave pipeline multimodal approach variant modalities k-means clustering multimodal emotion recognition system emotion recognition video-referenced system 3D CNN architectures pattern clustering audio–visual correlation audio-referenced system unimodal channels t-distributed stochastic neighbour embedding video signal processing convolutional neural nets modality-referenced system image fusion human emotion recognition master pipeline canonical correlation analysis correlation methods recognition performance audio-visual systems three-dimensional convolutional neural networks neural net architecture temporal data alignment temporal domain

来源：评论

学校读者我要写书评

暂无评论

SSGD: SPARSITY-PROMOTING stochastic GRADIENT DESCENT ALGORITHM FOR UNBIASED DNN PRUNING

SSGD: SPARSITY-PROMOTING STOCHASTIC GRADIENT DESCENT ALGORIT...

引用

IEEE International Conference on Acoustics, Speech, and signal processing (ICASSP)

作者： Lee, Ching-Hua Fedorov, Igor Rao, Bhaskar D. Garudadri, Harinath Univ Calif San Diego Dept ECE La Jolla CA 92093 USA ARM ML Res San Jose CA USA

ISBN: (纸本)9781509066315

While deep neural networks (DNNs) have achieved state-of-the-art results in many fields, they are typically over-parameterized. Parameter redundancy, in turn, leads to inefficiency. Sparse signal recovery (SSR) techniques, on the other hand, find compact solutions to overcomplete linear problems. Therefore, a logical step is to draw the connection between SSR and DNNs. In this paper, we explore the application of iterative reweighting methods popular in SSR to learning efficient DNNs. By efficient, we mean sparse networks that require less computation and storage than the original, dense network. We propose a reweighting framework to learn sparse connections within a given architecture without biasing the optimization process, by utilizing the affine scaling transformation strategy. The resulting algorithm, referred to as Sparsity-promoting stochastic Gradient Descent (SSGD), has simple gradient-based updates which can be easily implemented in existing deep learning libraries. We demonstrate the sparsification ability of SSGD on image classification tasks and show that it outperforms existing methods on the MNIST and CIFAR-10 datasets.

关键词： Deep learning sparse signal recovery network pruning iterative reweighting affine scaling

来源：评论

学校读者我要写书评

暂无评论

BigNeuron: a resource to benchmark and predict performance of algorithms for automated tracing of neurons in light microscopy datasets

引用

Nature methods 2023年第6期20卷 824-835页

作者： Linus Manubens-Gil Zhi Zhou Hanbo Chen Arvind Ramanathan Xiaoxiao Liu Yufeng Liu Alessandro Bria Todd Gillette Zongcai Ruan Jian Yang Miroslav Radojević Ting Zhao Li Cheng Lei Qu Siqi Liu Kristofer E Bouchard Lin Gu Weidong Cai Shuiwang Ji Badrinath Roysam Ching-Wei Wang Hongchuan Yu Amos Sironi Daniel Maxim Iascone Jie Zhou Erhan Bas Eduardo Conde-Sousa Paulo Aguiar Xiang Li Yujie Li Sumit Nanda Yuan Wang Leila Muresan Pascal Fua Bing Ye Hai-Yan He Jochen F Staiger Manuel Peter Daniel N Cox Michel Simonneau Marcel Oberlaender Gregory Jefferis Kei Ito Paloma Gonzalez-Bellido Jinhyun Kim Edwin Rubel Hollis T Cline Hongkui Zeng Aljoscha Nern Ann-Shyn Chiang Jianhua Yao Jane Roskams Rick Livesey Janine Stevens Tianming Liu Chinh Dang Yike Guo Ning Zhong Georgia Tourassi Sean Hill Michael Hawrylycz Christof Koch Erik Meijering Giorgio A Ascoli Hanchuan Peng Institute for Brain and Intelligence Southeast University Nanjing China. Microsoft Corporation Redmond WA USA. Tencent AI Lab Bellevue WA USA. Computing Environment and Life Sciences Directorate Argonne National Laboratory Lemont IL USA. Kaya Medical Seattle WA USA. University of Cassino and Southern Lazio Cassino Italy. Center for Neural Informatics Structures and Plasticity Krasnow Institute for Advanced Study George Mason University Fairfax VA USA. Faculty of Information Technology Beijing University of Technology Beijing China. Beijing International Collaboration Base on Brain Informatics and Wisdom Services Beijing China. Nuctech Netherlands Rotterdam the Netherlands. Janelia Research Campus Howard Hughes Medical Institute Ashburn VA USA. Department of Electrical and Computer Engineering University of Alberta Edmonton Alberta Canada. Ministry of Education Key Laboratory of Intelligent Computation and Signal Processing Anhui University Hefei China. Paige AI New York NY USA. Scientific Data Division and Biological Systems and Engineering Division Lawrence Berkeley National Lab Berkeley CA USA. Helen Wills Neuroscience Institute and Redwood Center for Theoretical Neuroscience UC Berkeley Berkeley CA USA. RIKEN AIP Tokyo Japan. Research Center for Advanced Science and Technology (RCAST) The University of Tokyo Tokyo Japan. School of Computer Science University of Sydney Sydney New South Wales Australia. Texas A&M University College Station TX USA. Cullen College of Engineering University of Houston Houston TX USA. Graduate Institute of Biomedical Engineering National Taiwan University of Science and Technology Taipei Taiwan. National Centre for Computer Animation Bournemouth University Poole UK. PROPHESEE Paris France. Department of Neuroscience Columbia University New York NY USA. Mortimer B. Zuckerman Mind Brain Behavior Institute Columbia University New York NY USA. Department of Computer Science Northern Illinois Universit

BigNeuron is an open community bench-testing platform with the goal of setting open standards for accurate and fast automatic neuron tracing. We gathered a diverse set of image volumes across several species that is representative of the data obtained in many neuroscience laboratories interested in neuron tracing. Here, we report generated gold standard manual annotations for a subset of the available imaging datasets and quantified tracing quality for 35 automatic tracing algorithms. The goal of generating such a hand-curated diverse dataset is to advance the development of tracing algorithms and enable generalizable benchmarking. Together with image quality features, we pooled the data in an interactive web application that enables users and developers to perform principal component analysis, t-distributed stochastic neighbor embedding, correlation and clustering, visualization of imaging and tracing data, and benchmarking of automatic tracing algorithms in user-defined data subsets. The image quality metrics explain most of the variance in the data, followed by neuromorphological features related to neuron size. We observed that diverse algorithms can provide complementary information to obtain accurate results and developed a method to iteratively combine methods and generate consensus reconstructions. The consensus trees obtained provide estimates of the neuron structure ground truth that typically outperform single algorithms in noisy datasets. However, specific algorithms may outperform the consensus tree strategy in specific imaging conditions. Finally, to aid users in predicting the most accurate automatic tracing results without manual annotations for comparison, we used support vector machine regression to predict reconstruction quality given an image volume and a set of automatic tracings.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A novel stochastic deep conviction network for emotion recognition in speech signal

引用

JOURNAL OF INTELLIGENT & FUZZY SYSTEMS 2020年第4期38卷 5175-5190页

作者： Shukla, Shilpi Jain, Madhu Mahatma Gandhi Miss Coll Engn & Technol Noida Uttar Pradesh India Jaypee Inst Informat Technol Dept Elect & Commun Engn Noida Uttar Pradesh India

Deep learning is far and wide considered to be the most powerful method in computer vision fields, which has a lot of applications such as image recognition, robot navigation systems, and self-driving cars. Recent developments in neural networks have led to an efficient end-to-end architecture to human activity representation and classification. In light of these recent events in deep learning, there is now much considerable concern about developing less expensive computation and memory-wise methods. This paper presents an optimized end-to-end approach named stochastic deep conviction network (SDCN) formulated using the deep learning method. It comprises of deep learning method namely deep belief network (DBN), two supervised machine learning algorithm support vector machine (SVM) and decision tree (DT) with optimization capability for speech emotion identification. In the beginning, pre-processing is performed and the features are automatically extracted from the input speech signal by the DBN. Since speech signal features loses most of the information and the performance cannot be guaranteed because dynamic interactions can generate uncountable emotion-specific experiences that have the same core feeling state but different perceptual inclinations so DBN provides more robust features. The next step is to classify the emotions in the training phase;here the SVM classifier is chosen which performs dual classification. In order to enhance this classification process, defects must be reduced and the best discrimination of the extracted features should be obtained hence particle swarm optimization (PSO) technique is being added along with SVM classifier in the training phase. To reduce the over fitting problem and risks of a single classifier a DT is being used in the testing phase for the exact identification of emotions (anger, disgust, fear, happiness, neutral and sadness) and therefore it obtains better performance than a single classifier. The complication of the

关键词： stochastic deep conviction network restricted Boltzmann machine particle swarm optimization support vector machine whale optimization

来源：评论

学校读者我要写书评

暂无评论

A Generalized stochastic Implementation of the Disparity Energy Model for Depth Perception

引用

JOURNAL OF signal processing SYSTEMS FOR signal image AND VIDEO TECHNOLOGY 2018年第5期90卷 709-725页

作者： Boga, Kaushik Leduc-Primeau, Francois Onizawa, Naoya Matsumiya, Kazumichi Hanyu, Takahiro Gross, Warren J. McGill Univ Dept Elect & Comp Engn Montreal PQ Canada Tohuku Univ Sendai Miyagi Japan

Implementing neuromorphic algorithms is increasingly interesting as the error resilience and low-area, low-energy nature of biological systems becomes the potential solution for problems in robotics and artificial intelligence. While conventional digital methods are inefficient in implementing massively parallel systems, analog solutions are hard to design and program. stochastic Computing (SC) is a natural bridge that allows pseudo-analog computations in the digital domain using low complexity hardware. However, large scale SC systems traditionally suffered from long latencies, hence higher energy consumption. This work develops a VLSI architecture for an SC based binocular vision system based on a disparity-energy model that emulates the hierarchical multi-layered neural structure in the primary visual cortex. The 3-layer neural network architecture is biologically plausible and is tuned to detecting 5 different disparities. The architecture is compact, adder-free, and achieves better disparity detection compared to a floating-point version by using a modified disparity-energy model. A generalized 1x100 pixel processing system is synthesized using TSMC 65nm CMOS technology and it achieves 71 % reduction in area-delay product and 48 % in energy savings compared to a fixed-point implementation at equivalent precision.

关键词： stochastic computing Neuromorphic computing Approximate computing Gabor filters Disparity-energy model computer vision Biomedical electronics neural networks

来源：评论

学校读者我要写书评

暂无评论

Saccade gaze prediction using a recurrent neural network 24

Saccade gaze prediction using a recurrent neural network

引用

24th IEEE International Conference on image processing, ICIP 2017

作者： Ngo, Thuyen Manjunath, B.S. Department of Electrical and Computer Engineering University of California Santa Barbara United States

ISBN: (纸本)9781509021758

We present a model that generates close-to-human gaze sequences for a given image in the free viewing task. The proposed approach leverages recent advances in image recognition using convolutional neural networks and sequence modeling with recurrent neural networks. Feature maps from convolutional neural networks are used as inputs to a recurrent neural network. The recurrent neural network acts like a visual working memory that integrates the scene information and outputs a sequence of saccades. The model is trained end-to-end with real-world human eye-tracking data using back propagation and adaptive stochastic gradient descent. Overall, the proposed model is simple compared to the state-of-the-art methods while offering favorable performance on a standard eye-tracking data set. © 2017 IEEE.

关键词： Eye tracking

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：