检索结果-内蒙古大学图书馆

arXiv 2020年

作者： Wong, Alexander Wang, Xiao Yu Hryniowski, Andrew Vision and Image Processing Research Group University of Waterloo WaterlooON Canada Waterloo Artificial Intelligence Institute University of Waterloo WaterlooON Canada DarwinAI Corp. WaterlooON Canada

A critical step to building trustworthy deep neural networks is trust quantification, where we ask the question: How much can we trust a deep neural network? In this study, we take a step towards simple, interpretable metrics for trust quantification by introducing a suite of metrics for assessing the overall trustworthiness of deep neural networks based on their behaviour when answering a set of questions. We conduct a thought experiment and explore two key questions about trust in relation to confidence: 1) How much trust do we have in actors who give wrong answers with great confidence?, and 2) How much trust do we have in actors who give right answers hesitantly? Based on insights gained, we introduce the concept of question-answer trust to quantify trustworthiness of an individual answer based on confident behaviour under correct and incorrect answer scenarios, and the concept of trust density to characterize the distribution of overall trust for an individual answer scenario. We further introduce the concept of trust spectrum for representing overall trust with respect to the spectrum of possible answer scenarios across correctly and incorrectly answered questions. Finally, we introduce NetTrustScore, a scalar metric summarizing overall trustworthiness. The suite of metrics aligns with past social psychology studies that study the relationship between trust and confidence. Leveraging these metrics, we quantify the trustworthiness of several well-known deep neural network architectures for image recognition to get a deeper understanding of where trust breaks down. The proposed metrics are by no means perfect, but the hope is to push the conversation towards better metrics to help guide practitioners and regulators in producing, deploying, and certifying deep learning solutions that can be trusted to operate in real-world, mission-critical scenarios. Copyright © 2020, The Authors. All rights reserved.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Insights into fairness through trust: Multi-scale trust quantification for financial deep learning

arXiv

引用

arXiv 2020年

作者： Wong, Alexander Hryniowski, Andrew Wang, Xiao Yu Vision and Image Processing Research Group University of Waterloo WaterlooON Canada Waterloo Artificial Intelligence Institute University of Waterloo WaterlooON Canada DarwinAI Corp. WaterlooON Canada

The success of deep learning in recent years have led to a significant increase in interest and prevalence for its adoption to tackle financial services tasks. One particular question that often arises as a barrier to adopting deep learning for financial services is whether the developed financial deep learning models are fair in their predictions, particularly in light of strong governance and regulatory compliance requirements in the financial services industry. A fundamental aspect of fairness that has not been explored in financial deep learning is the concept of trust, whose variations may point to an egocentric view of fairness and thus provide insights into the fairness of models. In this study we explore the feasibility and utility of a multi-scale trust quantification strategy to gain insights into the fairness of a financial deep learning model, particularly under different scenarios at different scales. More specifically, we conduct multi-scale trust quantification on a deep neural network for the purpose of credit card default prediction to study: 1) the overall trustworthiness of the model 2) the trust level under all possible prediction-truth relationships, 3) the trust level across the spectrum of possible predictions, 4) the trust level across different demographic groups (e.g., age, gender, and education), and 5) distribution of overall trust for an individual prediction scenario. The insights for this proof-of-concept study demonstrate that such a multi-scale trust quantification strategy may be helpful for data scientists and regulators in financial services as part of the verification and certification of financial deep learning solutions to gain insights into fairness and trust of these solutions. Copyright © 2020, The Authors. All rights reserved.

关键词： Finance

来源：评论

学校读者我要写书评

暂无评论

TinySpeech: Attention condensers for deep speech recognition neural networks on edge devices

arXiv

引用

arXiv 2020年

作者： Wong, Alexander Famouri, Mahmoud Pavlova, Maya Surana, Siddharth Vision and Image Processing Research Group University of Waterloo WaterlooON Canada Waterloo Artificial Intelligence Institute University of Waterloo WaterlooON Canada DarwinAI Corp. WaterlooON Canada

Advances in deep learning have led to state-of-the-art performance across a multitude of speech recognition tasks. Nevertheless, the widespread deployment of deep neural networks for on-device speech recognition remains a challenge, particularly in edge scenarios where the memory and computing resources are highly constrained (e.g., low-power embedded devices) or where the memory and computing budget dedicated to speech recognition is low (e.g., mobile devices performing numerous tasks besides speech recognition). In this study, we introduce the concept of attention condensers for building low-footprint, highly-efficient deep neural networks for on-device speech recognition on the edge. More specifically, an attention condenser is a self-attention mechanism that learns and produces a condensed embedding characterizing joint local and cross-channel activation relationships, and performs selective attention accordingly. Unlike self-attention mechanisms designed for deep convolutional neural networks that depend heavily on existing convolution modules, attention condensers act as self-contained, stand-alone modules and facilitate for efficient deep neural networks with much sparser use of larger stand-alone convolution modules and more frequent use of attention condensers. To illustrate its efficacy, we introduce TinySpeech, low-precision deep neural networks comprising largely of attention condensers tailored for on-device speech recognition using a machine-driven design exploration strategy. Experimental results on the Google Speech Commands benchmark dataset for limited-vocabulary speech recognition showed that TinySpeech networks achieved significantly lower architectural complexity (as much as 207× fewer parameters) and lower computational complexity (as much as 21× fewer multiply-add operations) when compared to previous deep neural networks in research literature. These results not only demonstrate the efficacy of attention condensers for building highly efficie

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Inter-layer information similarity assessment of deep neural networks via topological similarity and persistence analysis of data neighbour dynamics

arXiv

引用

arXiv 2020年

作者： Hryniowski, Andrew Wong, Alexander Vision and Image Processing Research Group University of Waterloo Canada Waterloo Artificial Intelligence Institute Canada DarwinAI Corp Canada

The quantitative analysis of information structure through a deep neural network (DNN) can unveil new insights into the theoretical performance of DNN architectures. Two very promising avenues of research towards quantitative information structure analysis are: 1) layer similarity (LS) strategies focused on the inter-layer feature similarity, and 2) intrinsic dimensionality (ID) strategies focused on layer-wise data dimensionality using pairwise information. Inspired by both LS and ID strategies for quantitative information structure analysis, we introduce two novel complimentary methods for inter-layer information similarity assessment premised on the interesting idea of studying a data sample’s neighbourhood dynamics as it traverses through a DNN. More specifically, we introduce the concept of Nearest Neighbour Topological Similarity (NNTS) for quantifying the information topology similarity between layers of a DNN. Furthermore, we introduce the concept of Nearest Neighbour Topological Persistence (NNTP) for quantifying the inter-layer persistence of data neighbourhood relationships throughout a DNN. The proposed strategies facilitate the efficient inter-layer information similarity assessment by leveraging only local topological information, and we demonstrate their efficacy in this study by performing analysis on a deep convolutional neural network architecture on image data to study the insights that can be gained with respect to the theoretical performance of a DNN. Copyright © 2020, The Authors. All rights reserved.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Fast Candidate Region Extraction for SAR Ship Target

Fast Candidate Region Extraction for SAR Ship Target

引用

Youth Academic Annual Conference of Chinese Association of Automation (YAC)

作者： Panpan Zhang Haibo Luo Zheng Xu Miao He Shenyang Institute of Automation Shenyang China Institutes for Robotics and Intelligent Manufacturing Shenyang China University of Chinese Academy of Sciences Beijing China Key Laboratory of Opto-Electronic Information Processing Shenyang China The Key Lab of Image Understanding and Computer Vision Shenyang China

ISBN: (纸本)9781665465373

At present, deep learning technology is widely used in ship target detection in synthetic aperture radar (SAR) images. However, high-resolution remote sensing SAR images cover a larger area and have larger image sizes. To be able to use the deep learning model for training and testing, the image needs to be cropped to the appropriate size. In a high-resolution SAR ship image, the ship target usually takes only a small part of the whole image. As a result, only part of the cropped image contains the target, and most of the rest are background regions. This will cause a lot of computational redundancy in the model inference stage. To solve this problem, a fast candidate region extraction algorithm is proposed in this paper for ship target extraction. The algorithm consists of three parts: firstly, extract the saliency map of the SAR ship image, secondly, perform median filtering on the segmented image, and thirdly, extract candidate regions. The superiority of the algorithm was demonstrated by experiments on the AIR-SARShip-1.0 dataset.

关键词： Deep learning Training Redundancy Radar polarimetry Inference algorithms Indexes Marine vehicles

来源：评论

学校读者我要写书评

暂无评论

A Deep Learning-Based Approach for Defect Detection and Removing on Archival Photos 18

A Deep Learning-Based Approach for Defect Detection and Remo...

引用

18th image processing: Algorithms and Systems Conference, IPAS 2020

作者： Sizyakin, R. Voronin, V. Gapon, N. Zelensky, A. Pižurica, A. Department Telecommunications and Information Processing IPITELIN-imec Ghent University Ghent Belgium Moscow State University of Technology "STANKIN" Moscow Russia Lab. "Mathematical methods of image processing and intelligent computer vision systems" Don State Technical University Rostov-on-Don Russia

Many archival photos are unique, existed only in a single copy. Some of them are damaged due to improper archiving (e.g. affected by direct sunlight, humidity, insects, etc.) or have physical damage resulting in the appearance of cracks, scratches on photographs, non-necessary signs, spots, dust, and so on. This paper proposed a system for detection and removing image defects based on machine learning. The method for detecting damage to an image consists of two main steps: the first step is to use morphological filtering as a pre-processing, the second step is to use the machine learning method, which is necessary to classify pixels that have received a massive response in the preprocessing phase. The second part of the proposed method is based on the use of the adversarial convolutional neural network for the reconstruction of damages detected at the previous stage. The effectiveness of the proposed method in comparison with traditional methods of defects detection and removal was confirmed experimentally. © 2020, Society for Imaging Science and Technology

关键词： Defects

来源：评论

学校读者我要写书评

暂无评论

TB-Net: A Tailored, Self-Attention Deep Convolutional Neural Network Design For Detection of Tuberculosis Cases From Chest X-Ray images

Research Square

引用

Research Square 2021年

作者： Wong, Alexander Hou Lee, James Ren Rahmat-Khah, Hadi Sabri, Ali Alaref, Amer Vision and Image Processing Research Group University of Waterloo Waterloo Canada Waterloo Artificial Intelligence Institute University of Waterloo Waterloo Canada DarwinAI Corp. Waterloo Canada Department of Radiology Niagara Health McMaster University Hamilton Canada Department of Diagnostic Radiology Thunder Bay Regional Health Sciences Centre Thunder Bay Canada Department of Diagnostic Imaging Northern Ontario School of Medicine Sudbury Canada

Background: Tuberculosis (TB) remains a global health problem, and is the leading cause of death from an infectious disease. A crucial step in the treatment of tuberculosis is screening high risk populations and the early detection of the disease, with chest x-ray (CXR) imaging being the most widely-used imaging modality. As such, there has been significant recent interest in artificial intelligence-based TB screening solutions for use in resource-limited scenarios where there is a lack of trained healthcare workers with expertise in CXR interpretation. Motivated by this pressing need and the recent recommendation by the World Health Organization (WHO) for the use of computer-aided diagnosis of TB in place of a human reader, we introduce TB-Net, a self-attention deep convolutional neural network tailored for TB case screening. Methods: We used CXR data from a multi-national patient cohort to train and test our models. A machine-driven design exploration approach leveraging generative synthesis was used to build a highly customized deep neural network architecture with attention condensers. Results: We conducted an explainability-driven performance validation process to validate TB-Net’s decision-making behaviour. Experiments on CXR data from a multi-national patient cohort showed that the proposed TB-Net is able to achieve accuracy/sensitivity/specificity of 99.86%/100.0%/99.71%. Radiologist validation was conducted on select cases by two board-certified radiologists with over 10 and 19 years of experience, respectively, and showed consistency between radiologist interpretation and critical factors leveraged by TB-Net for TB case detection for the case where radiologists identified anomalies. Conclusion: The proposed TB-Net not only achieves high tuberculosis case detection performance in terms of sensitivity and specificity, but also leverages clinically relevant critical factors in its decision making process. While not a production-ready solution, we hope that th

关键词： Computer aided diagnosis

来源：评论

学校读者我要写书评

暂无评论

COVIDNet-CT: A tailored deep convolutional neural network design for detection of COVID-19 cases from chest CT images

arXiv

引用

arXiv 2020年

作者： Gunraj, Hayden Wang, Linda Wong, Alexander Department of Mechanical and Mechatronics Engineering University of Waterloo Canada Vision and Image Processing Research Group University of Waterloo Canada DarwinAI Corp. Canada

The coronavirus disease 2019 (COVID-19) pandemic continues to have a tremendous impact on patients and healthcare systems around the world. In the fight against this novel disease, there is a pressing need for rapid and effective screening tools to identify patients infected with COVID-19, and to this end CT imaging has been proposed as one of the key screening methods which may be used as a complement to RT-PCR testing, particularly in situations where patients undergo routine CT scans for non-COVID-19 related reasons, patients with worsening respiratory status or developing complications that require expedited care, and patients suspected to be COVID-19-positive but have negative RT-PCR test results. Early studies on CT-based screening have reported abnormalities in chest CT images which are characteristic of COVID-19 infection, but these abnormalities may be difficult to distinguish from abnormalities caused by other lung conditions. Motivated by this, in this study we introduce COVIDNet-CT, a deep convolutional neural network architecture that is tailored for detection of COVID-19 cases from chest CT images via a machine-driven design exploration approach. Additionally, we introduce COVIDx-CT, a benchmark CT image dataset derived from CT imaging data collected by the China National Center for Bioinformation comprising 104,009 images across 1,489 patient cases. Furthermore, in the interest of reliability and transparency, we leverage an explainability-driven performance validation strategy to investigate the decision-making behaviour of COVIDNet-CT, and in doing so ensure that COVIDNet-CT makes predictions based on relevant indicators in CT images. Both COVIDNet-CT and the COVIDx-CT dataset are available to the general public in an open-source and open access manner as part of the COVID-Net initiative. While COVIDNet-CT is not yet a production-ready screening solution, we hope that releasing the model and dataset will encourage researchers, clinicians, and citize

关键词： Coronavirus

来源：评论

学校读者我要写书评

暂无评论

A True Global Contrast Method for IR Small Target Detection Under Complex Background

引用

IEEE Transactions on Geoscience and Remote Sensing 2025年 63卷

作者： Jinhui Han Saed Moradi Bo Zhou Wei Wang Qian Zhao Zhen Luo College of Physics and Telecommunication Engineering Zhoukou Normal University Zhoukou China Department of Systems Design Engineering Vision and Image Processing Laboratory University of Waterloo Waterloo ON Canada School of Electronic Information and Communications Huazhong University of Science and Technology Wuhan China School of Artificial Intelligence Zhoukou Normal University Zhoukou China

Local contrast methods are efficient approaches for infrared (IR) small target detection. However, since its baseline is obtained locally, a local contrast method can work well only when the target is locally prominent, which is not always the case when the background is very complex. In this article, a new detection framework named true global contrast measure (TGCM) is proposed, in which its baseline is obtained by global decomposition, so that it can handle situations where the target is not locally salient. For example, when it is adjacent to some bright background. The complex edges/corners and noises that are easily lost during decomposition will be suppressed through some simple but ingenious design. A Gaussian filtering operation is adopted to suppress noises, and a local max dilation is utilized on the separated baseline to maintain as much edge/corner information as possible. The residual information during decomposition will be used as a weighting function to further suppress clutters. Experiments on both real and simulated IR images demonstrated the effectiveness of the proposed method. Compared to some existing algorithms, it has advantages in target enhancement and background suppression. In particular, it can achieve good performance in certain scenarios where the target is not locally prominent.

关键词： Object detection Noise image edge detection Filtering Brightness Estimation Gradient methods Detectors Convolution Artificial intelligence

来源：评论

学校读者我要写书评

暂无评论

image saliency detection via multi-feature and manifold-space ranking 2021

Image saliency detection via multi-feature and manifold-spac...

引用

3rd Asia Pacific Information Technology Conference, APIT 2021

作者： Li, Xiaoli Zhao, Huaici Liu, Yunpeng Shenyang Institute of Automation Chinese Academy of Sciences Shenyang110016 China Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences Shenyang110169 China University of Chinese Academy of Sciences Beijing100049 China Key Laboratory of Opto-Electronic Information Processing Shenyang110016 China The Key Lab of Image Understanding and Computer Vision Shenyang110016 China Shenyang Jianzhu University Shenyang110168 China

ISBN: (纸本)9781450388108

In this paper, we propose an image saliency detection method by using multi-feature and manifold-space ranking. Basically, the proposed method extracts the color-histogram feature to obtain the fine information of the image, and the color-mean feature to obtain the coarse information respectively. To further improve the detection accuracy of the feature correlation between different image units, a manifold-space ranking method is used to calculate saliency values of image units to construct a saliency map on each feature-space. Specifically, we fuse the two saliency maps to obtain the final saliency map. Extensive experiments demonstrate that the proposed method not only outperforms the other methods, but also improves the accuracy and robustness of the saliency detection. © 2021 ACM.

关键词： Color

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：