检索结果-内蒙古大学图书馆

Artificial Intelligence (CAI), IEEE Conference on

作者： Junhao Dong Melvin Wong Sihan Xia Joel Wei En Tay College of Computing and Data Science (CCDS) Nanyang Technological University (NTU) Singapore Singapore Institute of Manufacturing Technology (SIMTech) A*STAR Singapore

ISBN: (数字)9798350354096

ISBN: (纸本)9798350354102

A well-established study of adversarial training has been proposed to improve network robustness against adversarial examples in the context of deep learning. However, its performance highly relies on large-scale training data. To relieve from such a data-hungry learning nature, we propose an efficient extension of adversarial training by conducting a data reduction method from a new perspective of generated data. The reduced dataset can be regarded as an alternative pre-training dataset, which promotes adversarial training methods for better robustness even than the original dataset. Experimental results and analyses demonstrate the effectiveness of our data reduction method, achieving the same level of adversarial robustness with a dataset volume reduced to 80% of its original size.

关键词： Training Deep learning Training data Learning (artificial intelligence) Robustness

来源：评论

学校读者我要写书评

暂无评论

Smartphone Fundus Photography Enhancement for Retinopathy of Prematurity Disease Diagnosis Using Deep Learning

Smartphone Fundus Photography Enhancement for Retinopathy of...

引用

International Conference of Computer and Informatics Engineering (IC2IE)

作者： Elizabeth Ndunge Mutua Bernard Shwabwabo Kasamani Christoph Reich School of Computing & Engineering Sciences Strathmore University Nairobi Kenya Institute for Data Science Cloud Computing & IT Society Furtwangen University Furtwangen Germany

Retinopathy of Prematurity (ROP) is the highest cause of childhood blindness globally with babies born preterm having a higher probability of contracting the disease. The disease diagnosis remains an economic burden to many countries, lack of enough ophthalmologists for the disease diagnosis coupled with non-existent national screening guidelines still remains a challenge. To diagnose the disease, a fundus photography is conducted, printout images are analyzed to determine the presence or absence of the disease. With the increase in the development of smartphones having advanced image capturing and processing features, the utilization of smartphones to capture retina image for disease diagnosis is becoming a common trend. For regions where ophthalmologists are few and/or for low resource regions with few or no retina capturing equipment, the use of smartphones to capture retina images for retina diseases is an effective method. This, however, is challenged: different smartphones produce images of different resolutions; some images are darker others lighter and with different resolution. A smartphone retina image capturing has a smaller field of view ranging between 450–900 which is a major limitation. A lens to support a bigger view can be combined with this approach to provide a wide view of 1300. This enlargement however distorts the image quality and may result in losing some image features. To overcome these challenges, this work develops an improved U-Net model to preprocess images captured using smartphones for ROP disease diagnosis. Our focus is to determine the presence or absence of the disease from smartphone captured images. Because the images are captured under a smaller field of view (FOV), we develop an improved U-Net model by adding patches to enhance image circumference and extract all features from the image and use the extracted features to train a U-Net model for the disease diagnosis. The model results outperformed similar recent developments wit

关键词：

来源：评论

学校读者我要写书评

暂无评论

Pose-Transformation and Radial Distance Clustering for Unsupervised Person Re-identification 32

Pose-Transformation and Radial Distance Clustering for Unsup...

引用

32nd British Machine Vision Conference, BMVC 2021

作者： Seth, Siddharth Sonth, Akash Chakraborty, Anirban Visual Computing Lab Department of Computational and Data Sciences Indian Institute of Science Bangalore India

Person re-identification (re-ID) aims to tackle the problem of matching identities across non-overlapping cameras. Supervised approaches require identity information that may be difficult to obtain and are inherently biased towards the dataset they are trained on, making them unscalable across domains. To overcome these challenges, we propose an unsupervised approach to the person re-ID setup. Having zero knowledge of true labels, our proposed method enhances the discriminating ability of the learned features via a novel two-stage training strategy. The first stage involves training a deep network on an expertly designed pose-transformed dataset obtained by generating multiple perturbations for each original image in the pose space. Next, the network learns to map similar features closer in the feature space using the proposed discriminative clustering algorithm. We introduce a novel radial distance loss, that attends to the fundamental aspects of feature learning - compact clusters with low intra-cluster and high inter-cluster variation. Extensive experiments on several large-scale re-ID datasets demonstrate the superiority of our method compared to state-of-the-art approaches. © 2021. The copyright of this document resides with its authors.

关键词： Clustering algorithms

来源：评论

学校读者我要写书评

暂无评论

Exploring Drug Repurposing Success Stories Through a Network-based Approach: Insights from a Case Study

Exploring Drug Repurposing Success Stories Through a Network...

引用

2024 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2024

作者： Pierini, Elisa Mazza, Ludovica Petti, Manuela Tieri, Paolo Sapienza University of Rome Data Science MSc Program Rome Italy Sapienza University of Rome Diag Dept. Comp. Contr. Manag. Eng. Rome Italy Cnr National Research Council Iac Institute for Applied Computing Rome Italy

ISBN: (纸本)9798350386226

Drug repositioning is a promising strategy to discover new therapeutic applications for existing drugs, significantly reducing the time and costs associated with traditional drug development. This study employs a network medicine approach to analyze successful cases of drug repositioning, focusing on the exploratory hypothesis that the efficacy of repositioning may be determined by functional similarity between between diseases for which the drug was originally designed and diseases for which the same drug is reused. Network medicine tools were employed to investigate the connections between disease-associated genes, proteins, and approved drugs. Biological networks, including protein-protein interactions and functional interactions networks, as well as gene- and drug-disease association data are analyzed to identify functional similarities and possible molecular connections between diseases and treatments. Using clustering techniques and topological analysis, the results reveal a suggestive overlap of involved genes and functional interactions, emphasizing the value of computational methods in accelerating drug repositioning efforts and improving understanding of drug repositioning dynamics for more efficient therapeutic interventions. © 2024 IEEE.

关键词： Network theory (graphs)

来源：评论

学校读者我要写书评

暂无评论

Visual Effects for Real Time Ocean Water Rendering 10

Visual Effects for Real Time Ocean Water Rendering

引用

10th International Congress on Advanced Applied Informatics, IIAI-AAI 2021

作者： Litvak, Leonard D. Deligiannidis, Leonidas School of Computing and Data Science Wentworth Institute of Technology BostonMA02115 United States

ISBN: (纸本)9781665424202

It is complex to visually represent the surface of ocean water because it is comprised of a multitude of visual effects such as transparency, translucency, and subsurface scattering. We implemented a diverse set of real time rendering techniques by which we achieved a fast and realistic looking ocean simulation. Though our combination of techniques, we are able to achieve realistic-looking ocean water while staying within the technical constraints of real-time rendering at 60 frames per second. To illustrate this, we created a GPU-accelerated rendered utilizing OpenGL containing an implementation of these techniques, combined, for the purpose of rendering ocean water. A full suite of systems was implemented for the ocean rendering engine, including waveform simulation, efficient projection of the ocean onto a perceived infinite grid, dynamic lighting, and all the associated rendering subsystems. Through our combination of our selected techniques, we were able to create a realistic approximation of an infinite ocean surface rendered in only several milliseconds per frame. © 2021 IEEE.

关键词： Rendering (computer graphics)

来源：评论

学校读者我要写书评

暂无评论

PERIODICITY DECOUPLING FRAMEWORK FOR LONG-TERM SERIES FORECASTING 12

PERIODICITY DECOUPLING FRAMEWORK FOR LONG-TERM SERIES FORECA...

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Dai, Tao Wu, Beiliang Liu, Peiyuan Li, Naiqi Bao, Jigang Jiang, Yong Xia, Shu-Tao College of Computer Science and Software Engineering Shenzhen University China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University China Tsinghua Shenzhen International Graduate School Tsinghua University Shenzhen China WeBank Institute of Financial Technology Shenzhen University China

Convolutional neural network (CNN)-based and Transformer-based methods have recently made significant strides in time series forecasting, which excel at modeling local temporal variations or capturing long-term dependencies. However, real-world time series usually contain intricate temporal patterns, thus making it challenging for existing methods that mainly focus on temporal variations modeling from the 1D time series directly. Based on the intrinsic periodicity of time series, we propose a novel Periodicity Decoupling Framework (PDF) to capture 2D temporal variations of decoupled series for long-term series forecasting. Our PDF mainly consists of three components: multi-periodic decoupling block (MDB), dual variations modeling block (DVMB), and variations aggregation block (VAB). Unlike the previous methods that model 1D temporal variations, our PDF mainly models 2D temporal variations, decoupled from 1D time series by MDB. After that, DVMB attempts to further capture short-term and long-term variations, followed by VAB to make final predictions. Extensive experimental results across seven real-world long-term time series datasets demonstrate the superiority of our method over other state-of-the-art methods, in terms of both forecasting performance and computational efficiency. Code is available at https://***/Hank0626/PDF. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Time series

来源：评论

学校读者我要写书评

暂无评论

Navigating Towards Fairness with data Selection 39

Navigating Towards Fairness with Data Selection

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Zhang, Yixuan Li, Zhidong Wang, Yang Chen, Fang Fan, Xuhui Zhou, Feng School of Statistics and Data Science Southeast University China Data Science Institute University of Technology Sydney Australia School of Computing Macquarie University Australia Center for Applied Statistics School of Statistics Renmin University of China China Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing China

ISBN: (纸本)157735897X

Machine learning algorithms often struggle to eliminate inherent data biases, particularly those arising from unreliable labels, which poses a significant challenge in ensuring fairness. Existing fairness techniques that address label bias typically involve modifying models and intervening in the training process, but these lack flexibility for large-scale datasets. To address this limitation, we introduce a data selection method designed to efficiently and flexibly mitigate label bias, tailored to more practical needs. Our approach utilizes a zero-shot predictor as a proxy model that simulates training on a clean holdout set. This strategy, supported by peer predictions, ensures the fairness of the proxy model and eliminates the need for an additional holdout set, which is a common requirement in previous methods. Without altering the classifier’s architecture, our modality-agnostic method effectively selects appropriate training data and has proven efficient and effective in handling label bias and improving fairness across diverse datasets in experimental evaluations. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Twitter Spam Detection Using Word Vector and Binary Classifier

Twitter Spam Detection Using Word Vector and Binary Classifi...

引用

2023 International Conference on Self Sustainable Artificial Intelligence Systems, ICSSAS 2023

作者： Marimuthu, Ramaprabha Muthulingam, Gurusigaamani Ayyanar Vinoth, N.A.S. Ganesan, Sivasathiya Ponnuchamy, Kardeepa Vel Tech Rangarajan Dr. Sagunthala R&d Institute of Science and Technology Department of Computer Science and Engineering Chennai Avadi India Kalasalingam Academy of Research and Education Department of Computer Science and Engineering Virudhunagar India Srm Institute of Science and Technology Department of Computing Technologies Chennai Kattankulathur India Easwari Engineering College Department of Artificial Intelligence and Data Science Chennai Ramapuram India

ISBN: (纸本)9798350300857

In recent years, social media communication plays an important role in sharing messages, thoughts, comments and ideas around the world. Among them, Twitter is one of the most popular communication mediums. Twitter provides free services for sharing information's that are limited to 140 characters. Each month nearly 40 million new twitter accounts were created. Since usage of twitter increases rapidly, fake accounts were created for posting the spam messages and links for redirecting to phishing websites. Researchers proposed various machine learning methodologies for filtering the spam to handle and maintain the social network security problems. But still detecting spam tweets, malicious link in real world scenario is still a complex task. In the proposed work, a novel deep learning technique is introduced to address the above mention issues. Word Vector model is used for learning the syntax of each tweet. Later by using proceeding representation data set, binary classifier is developed. For experimental purpose 1-month twitter datasets are collected. Proposed work outperforms when compared with various existing methodologies and also to detect the non-text based as well. The ratio of spam is determined using the precision, recall, and F-measure metrics. © 2023 IEEE.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

TOWARDS ROBUST OUT-OF-DISTRIBUTION GENERALIZATION BOUNDS VIA SHARPNESS

arXiv

引用

arXiv 2024年

作者： Zou, Yingtian Kawaguchi, Kenji Liu, Yingnan Liu, Jiashuo Lee, Mong-Li Hsu, Wynne School of Computing National University of Singapore Singapore Institute of Data Science National University of Singapore Singapore Department of Computer Science & Technology Tsinghua University China

Generalizing to out-of-distribution (OOD) data or unseen domain, termed OOD generalization, still lacks appropriate theoretical guarantees. Canonical OOD bounds focus on different distance measurements between source and target domains but fail to consider the optimization property of the learned model. As empirically shown in recent work, the sharpness of learned minima influences OOD generalization. To bridge this gap between optimization and OOD generalization, we study the effect of sharpness on how a model tolerates data change in domain shift which is usually captured by "robustness" in generalization. In this paper, we give a rigorous connection between sharpness and robustness, which gives better OOD guarantees for robust algorithms. It also provides a theoretical backing for "flat minima leads to better OOD generalization". Overall, we propose a sharpness-based OOD generalization bound by taking robustness into consideration, resulting in a tighter bound than non-robust guarantees. Our findings are supported by the experiments on a ridge regression model, as well as the experiments on deep learning classification tasks. Copyright © 2024, The Authors. All rights reserved.

关键词： Regression analysis

来源：评论

学校读者我要写书评

暂无评论

An Ensemble Deep Learning Model for Diabetic Retinopathy Identification 9

An Ensemble Deep Learning Model for Diabetic Retinopathy Ide...

引用

9th IEEE International Conference on Smart Structures and Systems, ICSSS 2023

作者： Hussain, Mohammad Manzoor Shanmugam, P. Moorthi, K. Sakthivelu, U. Rajasekar, A. Kumar, R. Naresh B v Raju Institute of Technology Department of Computer Science and Engineering Telangana Medak India Rajalakshmi Engineering College Department Computer Science and Engineering Chennai India Sairam Engineering College Department Electronics and Communication Engineering Chennai India Srm Institute of Science and Technology School of Computing Department of Networking and Communications Tamilnadu Kattankulathur India Sri Sairam Institute of Technology Department of Artificial Intelligence and Data Science Chennai India

ISBN: (纸本)9798350384208

Diabetic Retinopathy (DR) is an eye disease that can potentially cause significant injury in those diagnosed with diabetes. The condition arises due to issues with blood circulation to the retina, the delicate tissue in the posterior part of the eye. DR is recognized as the leading factor contributing to visual impairment among those affected by diabetes. However, timely identification and intervention might mitigate the occurrence of this condition. Conversely, the timely identification of DR poses challenges, while the diagnostic process can be laborious and time-intensive. The automated detection of DR necessitates categorizing colour retinal fundus pictures. This article presents a computer-aided diagnostic strategy that utilizes Deep Learning (DL) algorithms. In order to classify retinal fundus pictures, this study employed a combination of DL models, specifically ResNet50 and Inception V3. The studies are being conducted on the DIARETDB1 dataset. An analysis is conducted to compare the results of the suggested methodology. The comparison study results show that the proposed approach has more significant accuracy levels than the other methods. A high degree of accuracy, that is, 98.37%, has been attained by utilising the proposed techniques. © 2023 IEEE.

关键词： Deep learning

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：