检索结果-内蒙古大学图书馆

Anomaly detection of time series with smoothness-inducing sequential variational auto-encoder

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Li, Longyuan Yan, Junchi Wang, Haiyang Jin, Yaohui State Key Lab of Advanced Optical Communication System and Network MoE Key Lab of Artificial Intelligence AI Institute Shanghai Jiao Tong University Shanghai200240 China Department of Computer Science and Engineering MoE Key Lab of Artificial Intelligence AI Institute Shanghai Jiao Tong University Shanghai200240 China

Deep generative models have demonstrated their effectiveness in learning latent representation and modeling complex dependencies of time series. In this paper, we present a Smoothness-Inducing Sequential Variational Auto-Encoder (SISVAE) model for robust estimation and anomaly detection of multi-dimensional time series. Our model is based on Variational Auto-Encoder (VAE), and its backbone is fulfilled by a Recurrent Neural Network to capture latent temporal structures of time series for both generative model and inference model. Specifically, our model parameterizes mean and variance for each time-stamp with flexible neural networks, resulting in a non-stationary model that can work without the assumption of constant noise as commonly made by existing Markov models. However, such a flexibility may cause the model fragile to anomalies. To achieve robust density estimation which can also benefit detection tasks, we propose a smoothness-inducing prior over possible estimations. The proposed prior works as a regularizer that places penalty at non-smooth reconstructions. Our model is learned efficiently with a novel stochastic gradient variational Bayes estimator. In particular, we study two decision criteria for anomaly detection: reconstruction probability and reconstruction error. We show the effectiveness of our model on both synthetic datasets and public real-world benchmarks. © 2021, CC BY.

关键词： Recurrent neural networks

Data augmentation in microscopic images for material data mining

学校读者我要写书评

暂无评论

npj Computational Materials 2020年第1期6卷 601-609页

作者： Boyuan Ma Xiaoyan Wei Chuni Liu Xiaojuan Ban Haiyou Huang Hao Wang Weihua Xue Stephen Wu Mingfei Gao Qing Shen Michele Mukeshimana Adnan Omer Abuassba Haokai Shen Yanjing Su Beijing Advanced Innovation Center for Materials Genome Engineering University of Science and Technology BeijingBeijing 100083China School of Computer and Communication Engineering University of Science and Technology BeijingBeijing 100083China Beijing Key Laboratory of Knowledge Engineering for Materials Science Beijing 100083China Institute for Advanced Materials and Technology University of Science and Technology BeijingBeijing 100083China School of Materials Science and Engineering University of Science and Technology BeijingBeijing 100083China School of Materials Science and Technology Liaoning Technical UniversityLiaoning 114051China The Institute of Statistical Mathematics Research Organization of Information and SystemsTachikawaTokyo 190-8562Japan National Intellectual Property Administration Beijing 100088China Faculty of Engineering Sciences University of BurundiBujumburaBurundi Faculty of Engineering and Technology Palestine Technical University–KadoorieTulkaremPalestine College of Information Science and Engineering China University of PetroleumBeijingChina Key Lab of Petroleum Data Mining China University of PetroleumBeijingChina

Recent progress in material data mining has been driven by high-capacity models trained on large ***,collecting experimental data(real data)has been extremely costly owing to the amount of human effort and expertise ***,we develop a novel transfer learning strategy to address problems of small or insufficient *** strategy realizes the fusion of real and simulated data and the augmentation of training data in a data mining *** a specific task of grain instance image segmentation,this strategy aims to generate synthetic data by fusing the images obtained from simulating the physical mechanism of grain formation and the“image style”information in real *** results show that the model trained with the acquired synthetic data and only 35%of the real data can already achieve competitive segmentation performance of a model trained on all of the real *** the time required to perform grain simulation and to generate synthetic data are almost negligible as compared to the effort for obtaining real data,our proposed strategy is able to exploit the strong prediction power of deep learning without significantly increasing the experimental burden of training data preparation.

关键词： mining instance competitive

Rectangular unfoldings of polycubes 31

学校读者我要写书评

暂无评论

Rectangular unfoldings of polycubes

31st Canadian Conference on Computational Geometry, CCCG 2019

作者： Demaine, Martin L. Hearn, Robert Ku, Jason Uehara, Ryuhei Computer Science and Artificial Intelligence Lab MIT Electrical Engineering and Computer Science Department MIT Japan Advanced Institute of Science and Technology School of Information Science Japan

In this paper, we investigate the problem that asks if there exists a net of a polycube that is exactly a rectangle with slits. For this nontrivial question, we show affirmative solutions. First, we show some concrete examples: (1) no rectangle with slits with fewer than 24 squares can fold to any polycube, (2) a 4 × 7 rectangle with slits can fold to a heptacube (nonmanifold), (3) both of a 3 × 8 rectangle and a 4 × 6 rectangle can fold to a hexacube (nonmanifold), and (4) a 5 × 6 rectangle can fold to a heptacube (manifold). Second, we show a construction of infinite family of polycubes folded from a rectangle with slits. The smallest one given by this construction is a 6 × 20 rectangle with slits that can fold to a polycube of genus 5. This construction gives us a polycube for any positive genus. Moreover, by this construction, we can show that there exists a rectangle with slits that can fold to k different polycubes for any given positive integer k. © CCCG 2019. All rights reserved.

关键词：

Dual-Consistency semi-supervised learning with uncertainty quantification for covid-19 lesion segmentation from CT images

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Li, Yanwen Luo, Luyang Lin, Huangjing Chen, Hao Heng, Pheng-Ann Imsight AI Research Lab Shenzhen China Department of Computer Science and Engineering The Chinese University of Hong Kong Hong Kong Department of Computer Science and Engineering The Hong Kong University of Science and Technology Hong Kong Guangdong-Hong Kong-Macao Joint Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences China

The novel coronavirus disease 2019 (COVID-19) characterized by atypical pneumonia has caused millions of deaths worldwide. Automatically segmenting lesions from chest Computed Tomography (CT) is a promising way to assist doctors in COVID-19 screening, treatment planning, and follow-up monitoring. However, voxel-wise annotations are extremely expert-demanding and scarce, especially when it comes to novel diseases, while an abundance of unlabeled data could be available. To tackle the challenge of limited annotations, in this paper, we propose an uncertainty-guided dual-consistency learning network (UDC-Net) for semi-supervised COVID-19 lesion segmentation from CT images. Specifically, we present a dual-consistency learning scheme that simultaneously imposes image transformation equivalence and feature perturbation invariance to effectively harness the knowledge from unlabeled data. We then quantify the segmentation uncertainty in two forms and employ them together to guide the consistency regularization for more reliable unsupervised learning. Extensive experiments showed that our proposed UDC-Net improves the fully supervised method by 6.3% in Dice and outperforms other competitive semi-supervised approaches by significant margins, demonstrating high potential in real-world clinical practice. 2 © 2021, CC BY-NC-ND.

关键词： Coronavirus

OXnet: Deep omni-supervised thoracic disease detection from chest X-rays

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Luo, Luyang Chen, Hao Zhou, Yanning Lin, Huangjing Heng, Pheng-Ann Department of Computer Science and Engineering The Chinese University of Hong Kong Hong Kong Department of Computer Science and Engineering The Hong Kong University of Science and Technology Hong Kong Imsight AI Research Lab Shenzhen China Guangdong-Hong Kong-Macao Joint Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences China

Chest X-ray (CXR) is the most typical diagnostic X-ray examination for screening various thoracic diseases. Automatically localizing lesions from CXR is promising for alleviating radiologists' reading burden. However, CXR datasets are often with massive image-level annotations and scarce lesion-level annotations, and more often, without annotations. Thus far, unifying different supervision granularities to develop thoracic disease detection algorithms has not been comprehensively addressed. In this paper, we present OXnet, the first deep omni-supervised thoracic disease detection network to our best knowledge that uses as much available supervision as possible for CXR diagnosis. We first introduce supervised learning via a one-stage detection model. Then, we inject a global classification head to the detection model and propose dual attention alignment to guide the global gradient to the local detection branch, which enables learning lesion detection from image-level annotations. We also impose intra-class compactness and inter-class separability with global prototype alignment to further enhance the global information learning. Moreover, we leverage a soft focal loss to distill the soft pseudo-labels of unlabeled data generated by a teacher model. Extensive experiments on a large-scale chest X-ray dataset show the proposed OXnet outperforms competitive methods with significant margins. Further, we investigate omni-supervision under various annotation granularities and corroborate OXnet is a promising choice to mitigate the plight of annotation shortage for medical image diagnosis.2 © 2021, CC BY-NC-ND.

关键词： Deep learning

Learning to Know Where to See: A Visibility-Aware Approach for Occluded Person Re-identification

学校读者我要写书评

暂无评论

Learning to Know Where to See: A Visibility-Aware Approach f...

International Conference on computer Vision (ICCV)

作者： Jinrui Yang Jiawei Zhang Fufu Yu Xinyang Jiang Mengdan Zhang Xing Sun Yingcong Chen Wei-Shi Zheng School of Computer Science and Engineering Sun Yat-sen University China Pazhou Lab Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education China Youtu Lab Tencent Microsoft Research Asia The Hong Kong University of Science and Technology The Hong Kong University of Science and Technology (Guangzhou)

ISBN: (纸本)9781665428132

Person re-identification (ReID) has gained an impressive progress in recent years. However, the occlusion is still a common and challenging problem for recent ReID methods. Several mainstream methods utilize extra cues (e.g., human pose information) to distinguish human parts from obstacles to alleviate the occlusion problem. Although achieving inspiring progress, these methods severely rely on the fine-grained extra cues, and are sensitive to the estimation error in the extra cues. In this paper, we show that existing methods may degrade if the extra information is sparse or noisy. Thus we propose a simple yet effective method that is robust to sparse and noisy pose information. This is achieved by discretizing pose information to the visibility label of body parts, so as to suppress the influence of occluded regions. We show in our experiments that leveraging pose information in this way is more effective and robust. Besides, our method can be embedded into most person ReID models easily. Extensive experiments validate the effectiveness of our model on common occluded person ReID datasets.

关键词： Estimation error computer vision Annotations Computational modeling Robustness Noise measurement

Self-slimmed Vision Transformer

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Zong, Zhuofan Li, Kunchang Song, Guanglu Wang, Yali Qiao, Yu Leng, Biao Liu, Yu School of Computer Science and Engineering Beihang University China SenseTime Research China ShenZhen Key Lab of Computer Vision and Pattern Recognition SIAT-SenseTime Joint Lab Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences China University of Chinese Academy of Sciences China SIAT Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society China Shanghai AI Laboratory China

Vision transformers (ViTs) have become the popular structures and outperformed convolutional neural networks (CNNs) on various vision tasks. However, such powerful transformers bring a huge computation burden, because of the exhausting token-to-token comparison. The previous works focus on dropping insignificant tokens to reduce the computational cost of ViTs. But when the dropping ratio increases, this hard manner will inevitably discard the vital tokens, which limits its efficiency. To solve the issue, we propose a generic self-slimmed learning approach for vanilla ViTs, namely SiT. Specifically, we first design a novel Token Slimming Module (TSM), which can boost the inference efficiency of ViTs by dynamic token aggregation. As a general method of token hard dropping, our TSM softly integrates redundant tokens into fewer informative ones. It can dynamically zoom visual attention without cutting off discriminative token relations in the images, even with a high slimming ratio. Furthermore, we introduce a concise Feature Recalibration Distillation (FRD) framework, wherein we design a reverse version of TSM (RTSM) to recalibrate the unstructured token in a flexible auto-encoder manner. Due to the similar structure between teacher and student, our FRD can effectively leverage structure knowledge for better convergence. Finally, we conduct extensive experiments to evaluate our SiT. It demonstrates that our method can speed up ViTs by 1.7× with negligible accuracy drop, and even speed up ViTs by 3.6× while maintaining 97% of their performance. Surprisingly, by simply arming LV-ViT with our SiT, we achieve new state-of-the-art performance on ImageNet. Code is available at https://***/Sense-X/SiT. © 2021, CC BY.

关键词： Distillation

DiaBD: A Diabetes Dataset for Enhanced Risk Analysis and Research in Bangladesh

学校读者我要写书评

暂无评论

Data in Brief 2025年

作者： Tabia Tanzin Prama Md. Jobayer Rahman Marzia Zaman Farhana Sarker Khondaker A. Mamun Advanced Intelligent Multidisciplinary Systems (AIMS) Lab Institute of Research Innovation Incubation and Commercialization (IRIIC) United International University Dhaka 1212 Bangladesh CMED Health Ltd. Mohakhali DOHS Dhaka 1206 Bangladesh Department of Computer Science and Engineering Independent University Bangladesh (IUB) Dhaka 1229 Bangladesh Department of Computer Science and Engineering United International University (UIU) Dhaka 1212 Bangladesh

Diabetes is a chronic condition affecting millions worldwide and severely impacts health and quality of life. According to the International Diabetes Federation (IDF), over 463 million adults, which is 9.3% of the global population, live with diabetes. Diabetes ranks among the most prevalent chronic diseases and was the ninth-leading cause of mortality in 2019, with 4.2 million deaths reported. This article presents a dataset of 5,288 patient records from Bangladesh, addressing gaps in diabetes research and aiding in healthcare planning, risk analysis, and predictive modelling. The dataset comprises 14 attributes: age, gender, pulse rate, systolic and diastolic blood pressure, glucose levels, height, weight, body mass index (BMI), family history of diabetes and hypertension, cardiovascular disease (CVD), and stroke. A dependent attribute, Diabetic, indicates whether an individual has diabetes or not. The dataset ensures demographic diversity and precise measurements, supporting the study of diabetes and its related health issues. Features like CVD and stroke enable broader research on comorbidities. This dataset facilitates machine learning applications, risk assessment, and personalized healthcare strategies. Researchers can explore the links between diabetes, hypertension, CVD, and stroke, while healthcare providers and policymakers can leverage this to identify trends, allocate resources efficiently, and enhance public health strategies.

关键词： Diabetes Medical care Biometric Data Standardization Risk Factor Analysis Predictive Analysis

RCAnalyzer:visual analytics of rare categories in dynamic networks

学校读者我要写书评

暂无评论

Frontiers of Information Technology & Electronic engineering 2020年第4期21卷 491-506页

作者： Jia-cheng PAN Dong-ming HAN Fang-zhou GUO Da-wei ZHOU Nan CAO Jing-rui HE Ming-liang XU Wei CHEN State Key Lab of CAD&CG Zhejiang UniversityHangzhou 310058China Zhejiang Lab Hangzhou 311100China Department of Computer Science and Engineering Arizona State UniversityTempe 85281USA Intelligent Big Data Visualisation Lab Tongji UniversityShanghai 200082China School of Information Engineering Zhengzhou UniversityZhengzhou 450001China Henan Institute of Advanced Technology Zhengzhou UniversityZhengzhou 450001China

A dynamic network refers to a graph structure whose nodes and/or links dynamically change over *** visualization and analysis techniques focus mainly on summarizing and revealing the primary evolution patterns of the network *** work focuses on detecting anomalous changing patterns in the dynamic network,the rare occurrence of which could damage the development of the entire *** this study,we introduce the first visual analysis system RCAnalyzer designed for detecting rare changes of sub-structures in a dynamic *** proposed system employs a rare category detection algorithm to identify anomalous changing structures and visualize them in the context to help oracles examine the analysis results and label the *** particular,a novel visualization is introduced,which represents the snapshots of a dynamic network in a series of connected triangular *** clustering and optimal tree cut are performed on each matrix to illustrate the detected rare change of nodes and links in the context of their surrounding *** evaluate our technique via a case study and a user *** evaluation results verify the effectiveness of our system.

关键词： Rare category detection Dynamic network Visual analytics