检索结果-内蒙古大学图书馆

38th Conference on Neural information Processing Systems, NeurIPS 2024

作者： Saberi, Amirhossein Najafi, Amir Behjati, Amin Emrani, Ala Zolfi, Yasaman Shadrooy, Mahdi Motahari, Abolfazl Khalaj, Babak H. Department of Electrical Engineering Sharif University of Technology Tehran Iran Department of Computer Engineering Sharif University of Technology Tehran Iran Sharif Center for Information Systems and Data Science Sharif University of Technology Tehran Iran

The aim of this paper is to address the challenge of gradual domain adaptation within a class of manifold-constrained data distributions. In particular, we consider a sequence of T ≥ 2 data distributions P1, ..., PT undergoing a gradual shift, where each pair of consecutive measures Pi, Pi+1 are close to each other in Wasserstein distance. We have a supervised dataset of size n sampled from P0, while for the subsequent distributions in the sequence, only unlabeled i.i.d. samples are available. Moreover, we assume that all distributions exhibit a known favorable attribute, such as (but not limited to) having intra-class soft/hard margins. In this context, we propose a methodology rooted in Distributionally Robust Optimization (DRO) with an adaptive Wasserstein radius. We theoretically show that this method guarantees the classification error across all Pis can be suitably bounded. Our bounds rely on a newly introduced compatibility measure, which fully characterizes the error propagation dynamics along the sequence. Specifically, for inadequately constrained distributions, the error can exponentially escalate as we progress through the gradual shifts. Conversely, for appropriately constrained distributions, the error can be demonstrated to be linear or even entirely eradicated. We have substantiated our theoretical findings through several experimental results. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Breast Cancer Detection using Domain-Adversarial Training (DANN) with Invariant Risk Minimization (IRM) Hybrid Approach 3

Breast Cancer Detection using Domain-Adversarial Training (D...

引用

3rd International Conference on Electronics and Renewable Systems, ICEARS 2025

作者： Akash, K.V. Daniel, Esther Seetha, S. Durga, S. Karunya Institute of Technology and Sciences Division of Computer Science and Engineering Coimbatore India Cmr Institute of Technology Department of Information Science & Engineering India Amrita Vishwa Vidyapeetham Tifac Core in Cyber Security Amrita School of Engineering India

ISBN: (纸本)9798331509675

The breast cancer detection performs a key function in the health care network. The precise and early detection of cancer in the breast could aid to save life of the sufferer. The traditional machine learning methods struggle due to the data used to train is different from data used later this is known as Domain shifts. This project uses a hybrid model known as Domain-Adversarial Training of Neural Networks (DANN) with Invariant Risk Minimization (IRM) Hybrid Method to identification of the tumor in the breast. Through integrating the (DANN) and IRM works perfectly among the various type of data and also with the various image from the various sources. This model is more precise and reliable than any other older *** that this model act as an effective tool for detecting the breast cancer in the early stages. © 2025 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

A Transformer-Based Approach to Tracking and Counting Vessels in Shore-Side Surveillance Videos

A Transformer-Based Approach to Tracking and Counting Vessel...

引用

2023 International Conference on Machine Learning and Cybernetics, ICMLC 2023

作者： Fahn, Chin-Shyurng Zheng, Shan-Qian National Taiwan University of Science and Technology Department of Computer Science and Information Engineering Taipei106335 Taiwan

ISBN: (纸本)9798350303780

In this paper, we propose a transformer-based approach to tracking and counting vessels in shore-side surveillance videos. First, we detect the current frame of the video through a Visual Transformer to acquire vessel detection boxes and tracking boxes in the frame. We then perform a three-stage IoU matching using the transformer predictions. The first stage of IoU matching is carried out based on vessel appearance information. The second stage will retrieve the vessels that may be affected by the external environment from the frames of detection boxes with low confidence scores for IoU matching. And the third stage further executes IoU matching based on the motion information provided by the Kalman filter. In experiments, we compare the performance of our vessel tracking model with that of three classical models. According to the experimental results, our model can reach the precision, recall, multiple object tracking accuracy (MOTA), and the IDF1 of 92.4%, 68.1%, 62.5%, 75.1%, respectively, which performs the best on the whole. © 2023 IEEE.

关键词： Security systems

来源：评论

学校读者我要写书评

暂无评论

Audiovisual speech recognition based on a deep convolutional neural network

引用

Data Science and Management 2024年第1期7卷 25-34页

作者： Shashidhar Rudregowda Sudarshan Patilkulkarni Vinayakumar Ravi Gururaj H.L. Moez Krichen Department of Electronics and Communication Engineering JSS Science and Technology UniversityMysuru570006India Center for Artificial Intelligence Prince Mohammad Bin Fahd UniversityKhobar34754Saudi Arabia Department of Information Technology Manipal Institute of Technology BengaluruManipal Academy of Higher EducationManipal560064India Department of Information Technology Faculty of Computer Science and Information Technology(FCSIT)Al-Baha UniversityAlaqiq65779-7738Saudi Arabia ReDCAD Laboratory University of SfaxSfax3038Tunisia

Audiovisual speech recognition is an emerging research *** is the recognition of what someone is saying using visual information,primarily lip *** this study,we created a custom dataset for Indian English linguistics and categorized it into three main categories:(1)audio recognition,(2)visual feature extraction,and(3)combined audio and visual *** features were extracted using the mel-frequency cepstral coefficient,and classification was performed using a one-dimension convolutional neural *** feature extraction uses Dlib and then classifies visual speech using a long short-term memory type of recurrent neural ***,integration was performed using a deep convolutional *** audio speech of Indian English was successfully recognized with accuracies of 93.67%and 91.53%,respectively,using testing data from 200 *** training accuracy for visual speech recognition using the Indian English dataset was 77.48%and the test accuracy was 76.19%using 60 *** integration,the accuracies of audiovisual speech recognition using the Indian English dataset for training and testing were 94.67%and 91.75%,respectively.

关键词： Audiovisual speech recognition,Custom dataset 1D Convolution neural network(CNN) Deep CNN(DCNN) Long short-term memory(LSTM) Lipreading Dlib Mel-frequency cepstral coefficient(MFCC)

来源：评论

学校读者我要写书评

暂无评论

Ensuring Electronic Health Record (EHR) Privacy using Zero Knowledge Proofs (ZKP) and Secure Encryption Schemes on Blockchain 5

Ensuring Electronic Health Record (EHR) Privacy using Zero K...

引用

5th International Conference on Advancements in Computing, ICAC 2023

作者： Ranaweera, T.A.V.Y. Hewage, H.N.H. Hapuhinna, H.K.D.W.M.C.B. Preethilal, K.L.K.T. Senarathne, Amila Ruggahakotuwa, Laneesha Sri Lanka Institute of Information Technology Dept. Computer Systems Engineering Malabe Sri Lanka

ISBN: (纸本)9798350358131

In the haste of EHR digitization, the protection of patient data becomes paramount, casting a spotlight on small healthcare clinics grappling with the threats of an ever more interconnected healthcare sector. That is due to the existing centralized system and traditional password-based authentications which cause a single point of failure. Besides this patient consent for EHR sharing is narrowed to a one-time consent. To overcome these problems, this study suggests a robust solution that makes data sharing utilizing blockchain while achieving privacy using web-based Zero Knowledge Proofs (ZKP) authentication and Homomorphic Encryption (HE). The research also considers securely outsourcing EHRs to research organizations to improve data exchange for research purposes. Along with patient-controlled encryption, consent control, and time-limited data access, it also explores dynamic consent management while complying with healthcare standards. Lastly, this study stream-lines secure data sharing with insurance providers while ensuring patient privacy and defending against insurance fraud. © 2023 IEEE.

关键词： Blockchain

来源：评论

学校读者我要写书评

暂无评论

Parameters Estimation of Modified Triple Diode Model of PSCs Considering Charge Accumulations and Electric Field Effects Using Puma Optimizer

引用

computer Modeling in engineering & Sciences 2025年第4期143卷 723-745页

作者： Amlak Abaza Ragab A.El-Sehiemy Mona Gafar Ahmed Bayoumi Electrical Engineering Department Faculty of EngineeringKafrelsheikh UniversityKafrelsheikh33516Egypt Department of Computer Engineering and Information College of Engineering inWadi AlddawasirPrince Sattam Bin Abdulaziz UniversityKharj16278Saudi Arabia Machine Learning and Information Retrieval Department Artificial IntelligenceKafrelsheikh UniversityKafrelsheikh33516Egypt Electrical Department Faculty of Engineering and TechnologyBadr University in CairoBadr11829Egypt Physics and Mathematics Department Faculty of EngineeringKafrelsheikh UniversityKafrelsheikh33516Egypt College of Engineering Al-Ayen UniversityThi-Qar64001Iraq

Promoting the high penetration of renewable energies like photovoltaic(PV)systems has become an urgent issue for expanding modern power grids and has accomplished several challenges compared to existing distribution *** study measures the effectiveness of the Puma optimizer(PO)algorithm in parameter estimation of PSC(perovskite solar cells)dynamic models with hysteresis consideration considering the electric field effects on *** models used in this study will incorporate hysteresis effects to capture the time-dependent behavior of PSCs *** PO optimizes the proposed modified triple diode model(TDM)with a variable voltage capacitor and resistances(VVCARs)considering the hysteresis *** suggested PO algorithm contrasts with other wellknown optimizers from the literature to demonstrate its *** results emphasize that the PO realizes a lower RMSE(Root mean square errors),which proves its capability and efficacy in parameter extraction for the *** statistical results emphasize the efficiency and supremacy of the proposed PO compared to the other well-known competing *** convergence rates show good,fast,and stable convergence rates with lower RMSE via PO compared to the other five competitive ***,the lowermean realized via the PO optimizer is illustrated by the box plot for all optimizers.

关键词： Dynamic model of PSCs puma optimizer parameter estimation triple diode model

来源：评论

学校读者我要写书评

暂无评论

Analyzing Voice Quality with Multi-Dimensional Voice Program for Disease Determination 6

Analyzing Voice Quality with Multi-Dimensional Voice Program...

引用

6th International Symposium on computer, Consumer and Control, IS3C 2023

作者： Liza, Rahmi Tsung, Chen-Kun National Chin-Yi University of Technology dept. Computer Science and Information Engineering Taichung Taiwan

ISBN: (纸本)9798350301953

Multi-Dimensional Voice Program (MDVP) parameters are very popular among physicians/clinicians to detect vocal pathologies and analyze various diseases of the vocal cords. In this paper, voice pathologies are automatically detected using the parameters of the MDVP. However, MDVP is commercial software, so this work is trying to build MDVP using Python to extract MDVP parameters useful for various experiments, automatic detection of sound pathologies, and automatic classification of voice recognition. This study evaluates MDVP parameters and applies the XGBoost model as a classification method to analyze and classify diseases. This work considers three sample data, polyps, nodules, and Reinke edema, popular in clinical vocal cords diseases, from Saarbruecken Voice Database (SVD) for data testing and training. Test results demonstrate the excellent ability of MDVP's parameter extraction to identify healthy voices and obtain accurate classification results to discriminate between healthy voices and pathological voices. The best overall accuracy is 98% using the XGBoost classifier. © 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

Investigating Time Series Clustering Algorithms for Hyper Spectral Image Analysis 5th

Investigating Time Series Clustering Algorithms for Hyper Sp...

引用

5th International Conference on Data Science, Machine Learning and Applications, ICDSMLA 2023

作者： Mahalakshmi, R. Agarwal, Trapty Kudari, Jayashree M. Mehra, Ritika School of Computer Science and Engineering Presidency University Karnataka Bengaluru India Maharishi School of Engineering & Technology Maharishi University of Information Technology Uttar Pradesh Lucknow India Karnataka Bangalore India Department of Computer Science & Engineering School of Engineering and Computing Dev Bhoomi Uttarakhand University Navgaon Manduwala Dehradun India

ISBN: (纸本)9789819780426

Time collection clustering algorithms (TSCAs) are of increasing significance within the evaluation and interpretation of hyper spectral image (HSI) facts. TSCAs are being used to become aware of spatiotemporal styles of spectral characteristics in HSI records by identifying areas of interest (ROIs) which can be associated with physical techniques and mapping the ones ROIs to clusters. One-of-a-kind clustering algorithms may be used to analyze HSI statistics and tease out diffused spatiotemporal modifications. This paper investigates the software of two exceptional TSCAs, particularly okay-method clustering and Self-Organizing Maps (SOMs) to HSI *** SOM set of rules changed into additionally capable of phase the hyperspectral facts into meaningful clusters primarily based on their spectral traits. The accuracy of the SOM results was also assessed. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Self organizing maps

来源：评论

学校读者我要写书评

暂无评论

Efficient Clustered Mean Forecasting for Time Series with Head-Based Aggregation 5th

Efficient Clustered Mean Forecasting for Time Series with He...

引用

5th International Conference on Data Science, Machine Learning and Applications, ICDSMLA 2023

作者： Anand, Piyush Ramesh, T. Agarwal, Trapty Preethi, D. Department of Computer Science and Engineering School of Engineering and Computing Dev Bhoomi Uttarakhand University Navgaon Manduwala Dehradun India School of Computer Science and Engineering Presidency University Karnataka Bengaluru India Maharishi School of Engineering and Technology Maharishi University of Information Technology Uttar Pradesh Luknow India Karnataka Bangalore India

ISBN: (纸本)9789819780426

Green Clustered mean Forecasting (ECMF) for Time series is a novel forecasting approach that is capable of efficaciously aggregating the underlying additives of a time collection through combining the top-primarily based Aggregation (HBA) technique alongside the mean Forecasting (MF) method. ECMF may be utilized to detect and become aware of companies of time collection intently associated with every other, therefore taking into account extra correct and reliable forecasts. By way of counting on HBA to discover any patterns in the information and then imposing MF to generate forecasts from the ones styles, ECMF can offer more accurate and reliable forecasts than the prevailing strategies. Moreover, this technique is computationally green, due to the hierarchical shape of the data, which reduces the quantity of forecasting experiments wanted. The ensuing forecasts are also easier to examine, as they may be generated from a standard head-based totally version. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

DEFAEK: Domain Effective Fast Adaptive Network for Face Anti-Spoofing

引用

Neural Networks 2023年 161卷 83-92页

作者： Lin, Jiun-Da Han, Yue-Hua Huang, Po-Han Tan, Julianne Chen, Jun-Cheng Tanveer, M. Hua, Kai-Lung Department of Computer Science and Information Engineering National Taiwan University of Science and Technology Taipei106335 Taiwan Research Center for Information Technology Innovation Academia Sinica Taipei115201 Taiwan Department of Mathematics Indian Institute of Technology Indore Simrol453552 India

Existing deep learning based face anti-spoofing (FAS) or deepfake detection approaches usually rely on large-scale datasets and powerful networks with significant amount of parameters to achieve satisfactory performance. However, these make them resource-heavy and unsuitable for handheld devices. Moreover, they are limited by the types of spoof in the dataset they train on and require considerable training time. To produce a robust FAS model, they need large datasets covering the widest variety of predefined presentation attacks possible. Testing on new or unseen attacks or environments generally results in poor performance. Ideally, the FAS model should learn discriminative features that can generalize well even on unseen spoof types. In this paper, we propose a fast learning approach called Domain Effective Fast Adaptive nEt-worK (DEFAEK), a face anti-spoofing approach based on the optimization-based meta-learning paradigm that effectively and quickly adapts to new tasks. DEFAEK treats differences in an environment as domains and simulates multiple domain shifts during training. To further improve the effectiveness and efficiency of meta-learning, we adopt the metric learning in the inner loop update with careful sample selection. With extensive experiments on the challenging CelebA-Spoof and FaceForensics++ datasets, the evaluation results show that DEFAEK can learn cues independent of the environment with good generalization capability. In addition, the resulting model is lightweight following the design principle of modern lightweight network architecture and still generalizes well on unseen classes. In addition, we also demonstrate our model's capabilities by comparing the numbers of parameters, FLOPS, and model performance with other state-of-the-art methods. © 2023 Elsevier Ltd

关键词： Large dataset

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：