检索结果-内蒙古大学图书馆

41st International Conference on Machine Learning, ICML 2024

作者： Yang, Yahong He, Juncai Department of Mathematics The Pennsylvania State University University Park State CollegePA16802 United States Computer Electrical and Mathematical Science and Engineering Division The King Abdullah University of Science and Technology Thuwal23955 Saudi Arabia

Constructing the architecture of a neural network is a challenging pursuit for the machine learning community, and the dilemma of whether to go deeper or wider remains a persistent question. This paper explores a comparison between deeper neural networks (DeNNs) with a flexible number of layers and wider neural networks (WeNNs) with limited hidden layers, focusing on their optimal generalization error in Sobolev losses. Analytical investigations reveal that the architecture of a neural network can be significantly influenced by various factors, including the number of sample points, parameters within the neural networks, and the regularity of the loss function. Specifically, a higher number of parameters tends to favor WeNNs, while an increased number of sample points and greater regularity in the loss function lean towards the adoption of DeNNs. We ultimately apply this theory to address partial differential equations using deep Ritz and physics-informed neural network (PINN) methods, guiding the design of neural networks. Copyright 2024 by the author(s)

关键词： Partial differential equations

来源：评论

学校读者我要写书评

暂无评论

Attack-free Evaluating and Enhancing Adversarial Robustness on Categorical Data 41

Attack-free Evaluating and Enhancing Adversarial Robustness ...

引用

41st International Conference on Machine Learning, ICML 2024

作者： Zhou, Yujun Han, Yufei Zhuang, Haomin Bao, Hongyan Zhang, Xiangliang Department of Computer Science and Engineering University of Notre Dame Notre DameIN United States INRIA France King Abdullah University of Science and Technology Thuwal Saudi Arabia

Research on adversarial robustness has predominantly focused on continuous inputs, leaving categorical inputs, especially tabular attributes, less *** echo this challenge, our work aims to evaluate and enhance the robustness of classification over categorical attributes against adversarial perturbations through efficient attack-free *** propose a robustness evaluation metric named Integrated Gradient-Smoothed Gradient (IGSG).It is designed to evaluate the attributional sensitivity of each feature and the decision boundary of the classifier, two aspects that significantly influence adversarial risk, according to our theoretical *** this metric, we develop an IGSG-based regularization to reduce adversarial risk by suppressing the sensitivity of categorical *** conduct extensive empirical studies over categorical datasets of various application *** results affirm the efficacy of both IGSG and IGSG-based ***, IGSG-based regularization surpasses the state-of-the-art robust training methods by a margin of approximately 0.4% to 12.2% on average in terms of adversarial accuracy, especially on high-dimension *** code is available at https://***/YujunZhou/IGSG. Copyright 2024 by the author(s)

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Lactate-mediated medium-chain fatty acid production from expired dairy and beverage waste

引用

Environmental science and Ecotechnology 2024年第5期21卷 278-286页

作者： Bin Bian Wenxiang Zhang Najiaowa Yu Wei Yang Jiajie Xu Bruce E.Logan Pascal E.Saikaly Water Desalination and Reuse Center(WDRC) King Abdullah University of Science and Technology(KAUST)Thuwal 23955-6900Kingdom of Saudi Arabia Environmental Science and Engineering Program Biological and Environmental Science and Engineering(BESE)DivisionKing Abdullah University of Science and Technology(KAUST)Thuwal 23955-6900Kingdom of Saudi Arabia Department of Civil and Environmental Engineering The Pennsylvania State UniversityUniversity ParkPA16802USA Research Centre of Ecology&Environment for Coastal Area and Deep Sea Southern Marine Science and Engineering Guangdong Laboratory(Guangzhou)Guangzhou511458China School of Marine Science Ningbo UniversityNingbo315211China

Fruits,vegetables,and dairy products are typically the primary sources of household food ***,anaerobic digestion is the most used bioprocess for the treatment of food waste with concomitant generation of ***,to achieve a circular carbon economy,the organics in food waste should be converted to new chemicals with higher value than *** we demonstrate the feasibility of medium-chain carboxylic acid(MCCA)production from expired dairy and beverage waste via a chain elongation platform mediated by *** a two-stage fermentation process,the first stage with optimized operational conditions,including varying temperatures and organic loading rates,transformed expired dairy and beverage waste into lactate at a concentration higher than 900 mM C at 43℃.This lactate was then used to produce>500 mM C caproate and>300 mM C butyrate via microbial chain ***,lactate-producing microbes such as Lactobacillus and Lacticaseibacillus were regulated by temperature and could be highly enriched under mesophilic conditions in the firststage *** the second-stage chain elongation reactor,the dominating microbes were primarily from the genera Megasphaera and Caproiciproducens,shaped by varying feed and inoculum *** network analysis revealed positive correlations among species from the genera Caproiciproducens,Ruminococcus,and CAG-352,as well as Megasphaera,Bacteroides,and Solobacterium,indicating strong microbial interactions that enhance caproate *** findings suggest that producing MCCAs from expired dairy and beverage waste via lactate-mediated chain elongation is a viable method for sustainable waste management and could serve as a chemical production platform in the context of building a circular bioeconomy.

关键词： Microbial chain elongation Medium chain carboxylic acids Food waste Fermentation

来源：评论

学校读者我要写书评

暂无评论

On-Chip Quantum Dot Lasers Driven High-Speed Optical Neural Networks

On-Chip Quantum Dot Lasers Driven High-Speed Optical Neural ...

引用

CLEO: science and Innovations in CLEO 2024, CLEO: S and I 2024 - Part of Conference on Lasers and Electro-Optics

作者： Zhou, Zhican Fang, Yuetong Ou, Xiangpeng He, William Wu, Xuhao Xu, Renjing Pan, David Z. Wan, Yating Integrated Photonics Lab King Abdullah University of Science and Technology Thuwal Makkah Saudi Arabia Function Hub The Hong Kong University of Science and Technology Guangzhou China Department of Electrical and Computer Engineering The University of Texas Austin78703 United States

We present an optical neural network architecture driven by on-chip quantum dot lasers, achieving 3.5 TOPS/mm2 theoretical computational density. With an optimized model (reducing programming operations by over 75%), ... 详细信息

关键词： Quantum dot lasers

来源：评论

学校读者我要写书评

暂无评论

On-Chip Quantum Dot Lasers Driven High-Speed Optical Neural Networks

On-Chip Quantum Dot Lasers Driven High-Speed Optical Neural ...

引用

2024 Conference on Lasers and Electro-Optics, CLEO 2024

作者： Zhou, Zhican Fang, Yuetong Ou, Xiangpeng He, William Wu, Xuhao Xu, Renjing Pan, David Z. Wan, Yating Integrated Photonics lab King Abdullah University of Science and Technology Makkah Thuwal Saudi Arabia Function Hub The Hong Kong University of Science and Technology Guangzhou China Department of Electrical and Computer Engineering The University of Texas Austin78703 United States

ISBN: (纸本)9781957171395

关键词： Quantum dot lasers

来源：评论

学校读者我要写书评

暂无评论

ResMHA-Net:Enhancing Glioma Segmentation and Survival Prediction Using a Novel Deep Learning Framework

引用

computers, Materials & Continua 2024年第10期81卷 885-909页

作者： Novsheena Rasool Javaid Iqbal Bhat Najib Ben Aoun abdullah Alharthi Niyaz Ahmad Wani Vikram Chopra Muhammad Shahid Anwar Department of Computer Science School of Engineering and TechnologyIslamic University of Science and TechnologyAwantipora192122India Faculty of Computing and Technology Al-Baha UniversityAlaqiq65779-7738Saudi Arabia Research Groups in Intelligent Machines National School of Engineers of SfaxUniversity of SfaxSfax3038Tunisia Department of Electrical Engineering College of EngineeringKing Khalid UniversityAbha61421Saudi Arabia Department of Computer Science and Engineering Thapar Institute of Engineering and TechnologyPatiala147004Punjab Department of Electrical and Instrumentation Engineering Thapar Institute of Engineering and TechnologyPatiala147004Punjab Department of AI and Software Gachon UniversitySeongnam-si13120Republic of Korea

Gliomas are aggressive brain tumors known for their heterogeneity,unclear borders,and diverse locations on Magnetic Resonance Imaging(MRI)*** factors present significant challenges for MRI-based segmentation,a crucial step for effective treatment planning and monitoring of glioma *** study proposes a novel deep learning framework,ResNet Multi-Head Attention U-Net(ResMHA-Net),to address these challenges and enhance glioma segmentation ***-Net leverages the strengths of both residual blocks from the ResNet architecture and multi-head attention *** powerful combination empowers the network to prioritize informative regions within the 3D MRI data and capture long-range *** doing so,ResMHANet effectively segments intricate glioma sub-regions and reduces the impact of uncertain tumor *** rigorously trained and validated ResMHA-Net on the BraTS 2018,2019,2020 and 2021 ***,ResMHA-Net achieved superior segmentation accuracy on the BraTS 2021 dataset compared to the previous years,demonstrating its remarkable adaptability and robustness across diverse ***,we collected the predicted masks obtained from three datasets to enhance survival prediction,effectively augmenting the dataset *** features were then extracted from these predicted masks and,along with clinical data,were used to train a novel ensemble learning-based machine learning model for survival *** model employs a voting mechanism aggregating predictions from multiple models,leading to significant improvements over existing *** ensemble approach capitalizes on the strengths of various models,resulting in more accurate and reliable predictions for patient ***,we achieved an impressive accuracy of 73%for overall survival(OS)prediction.

关键词： Glioma MRI segmentation multihead attention survival prediction deep learning

来源：评论

学校读者我要写书评

暂无评论

Cost-Effective Authenticated Solution (CAS) for 6G-Enabled Artificial Intelligence of Medical Things (AIoMT)

引用

IEEE Internet of Things Journal 2024年第13期11卷 23977-23984页

作者： Mahmood, Khalid Obaidat, Mohammad S. Shamshad, Salman Alenazi, Mohammed J. F. Kumar, Gulshan Anisi, Mohammad Hossein Conti, Mauro National Yunlin University of Science and Technology Graduate School of Intelligent Data Science Douliu64002 Taiwan The University of Jordan King Abdullah II School of Information Technology Amman11942 Jordan University of Science and Technology Beijing School of Communication and Computing Engineering Beijing100083 China Amity University Amity School of Engineering and Technology Noida201301 India University of Lahore Department of Software Engineering Lahore54590 Pakistan King Saud University College of Computer and Information Sciences Department of Computer Engineering Riyadh12311 Saudi Arabia University of Padua Department of Mathematics Padua35131 Italy Lovely Professional University Department of Computer Science and Engineering Phagwara144411 India University of Essex School of Computer Science and Electronic Engineering ColchesterCO4 3SQ United Kingdom

The Internet of Things (IoT) is a network of interconnected objects, which congregate and exchange gigantic amounts of data. Usually, predeployed embedded sensors sense this massive data. Soon, several applications of IoT are anticipated to exploit emerging 6G technology. Healthcare is one of them, where the 6G-inspired paradigm may facilitate the users to exchange information through hundreds of sensors under the assumption of Artificial Intelligence of Things (AIoT). Integration of medical sensors with AIoT is known as Artificial Intelligence of Medical Things (AIoMT). The secure and seamless interactions among 6G-enabled AIoMT users should be the primary challenge. Furthermore, resource-constrained wearable sensing devices, with their inability to execute complex security solutions, provide an ideal attraction for malicious entities to launch diverse attacks. These challenges have motivated us to design a cost-effective authenticated solution (CAS) for 6G-enabled AIoMT healthcare applications. Our CAS protocol not only prevents cyber threats like impersonation session key secrecy but it can also prevent physical threats like hardware tampering. We observe formal and informal security validations to endorse its robustness and effectiveness. Performance comparison reveals that the CAS protocol offers maximum security enrichment. Moreover, CAS is cost-effective as it has achieved 33% and 60% reduction in computation and communication overheads, respectively, compared to contemporary competing related protocols. © 2014 IEEE.

关键词： Authentication

来源：评论

学校读者我要写书评

暂无评论

Assessment of the wall effects on the cellular structure by combined planar laser-induced fluorescence and Rayleigh scattering

Assessment of the wall effects on the cellular structure by ...

引用

AIAA science and technology Forum and Exposition, AIAA SciTech Forum 2025

作者： Alicherif, Mhedine Chavez, Samir B. Rojas Chatelain, Karl P. Guiberti, Thibault F. Lacoste, Deanna A. PO box 4700 Thuwal23955-6900 Saudi Arabia Mechanical Engineering Program Physical Science and Engineering Division King Abdullah University of Science and Technology Thuwal23955-6900 Saudi Arabia

ISBN: (数字)9781624107238

ISBN: (纸本)9781624107238

This study investigates the structure of the detonation front in a stoichiometric H2-air mixture (2H2-O2-3.76N2), seeded with 2000,ppm of NO, at 25,kPa and 293,K, for two channel positions: centered and near thewalls (side). Acombination of NO-PLIF and Rayleigh scattering techniques was used to characterize key features of the detonation front, such as unburned gas pockets, transverse waves, cellular irregularities, and triple point trajectories. The irregular features were categorized into four distinct "events", and their occurrence probabilities were determined for each configuration based on samples of over one hundred detonation cells. The centered configuration showed fewer irregularities with lower probabilities for all events, indicating a regular structure. In contrast, the side configuration showed higher probabilities for all events, indicating a more irregular structure closer to the walls. With the same techniques a correlation between the formation of unburned gas pockets and a local thickening of the front was investigated. While the processes seems to be interconnected in the center of the channel this correlation is lost closer to the walls. The results suggest a strong dependence between the occurrence detonation irregularities and their position relative to the walls. © 2025, American Institute of Aeronautics and Astronautics Inc, AIAA. All rights reserved.

关键词： Rayleigh scattering

来源：评论

学校读者我要写书评

暂无评论

A Via-Less Fully Screen-Printed Reconfigurable Intelligent Surface for 5G Millimeter Wave Communication

A Via-Less Fully Screen-Printed Reconfigurable Intelligent S...

引用

2023 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio science Meeting, AP-S/URSI 2023

作者： Yang, Yiming Wang, Ruiqi Vaseem, Mohammad Makki, Behrooz Shamim, Atif King Abdullah University of Science and Technology Computer Electrical and Mathematical Sciences and Engineering Division Thuwal Saudi Arabia Ericsson Research Gothenburg417 56 Sweden

ISBN: (纸本)9781665442282

In this paper, we propose a via-less fully screen-printed reconfigurable intelligent surface which can establish a second line-of-sight communication from 23.5GHz to 29.5GHz. By serially connecting the H shaped resonator along the H field of the incident wave, we minimize the effect of the biasing lines and make a via-less design, which reduces the fabrication difficulty and cost. The unit-cell simulation of the array with screen-printed VO2 switches shows a 215° to 160° phase shift difference between the ON and OFF states within bandwidth. During the field testing of the ideal arrays, we verify that the array can redirect the 45° incident wave to 0° reflection with a signal enhancement of at least 10 dB as compared to the array which has all unit cells in the OFF condition. © 2023 IEEE.

关键词： Vanadium dioxide

来源：评论

学校读者我要写书评

暂无评论

Understanding Forgetting in Continual Learning with Linear Regression 41

Understanding Forgetting in Continual Learning with Linear R...

引用

41st International Conference on Machine Learning, ICML 2024

作者： Ding, Meng Ji, Kaiyi Wang, Di Xu, Jinhui Department of Computer Science and Engineering State University of New York at Buffalo Buffalo United States Division of CEMSE King Abdullah University of Science and Technology Thuwal Saudi Arabia

Continual learning, focused on sequentially learning multiple tasks, has gained significant attention recently. Despite the tremendous progress made in the past, the theoretical understanding, especially factors contributing to catastrophic forgetting, remains relatively unexplored. In this paper, we provide a general theoretical analysis of forgetting in the linear regression model via Stochastic Gradient Descent (SGD) applicable to both under-parameterized and overparameterized regimes. Our theoretical framework reveals some interesting insights into the intricate relationship between task sequence and algorithmic parameters, an aspect not fully captured in previous studies due to their restrictive assumptions. Specifically, we demonstrate that, given a sufficiently large data size, the arrangement of tasks in a sequence-where tasks with larger eigenvalues in their population data covariance matrices are trained later-tends to result in increased forgetting. Additionally, our findings highlight that an appropriate choice of step size will help mitigate forgetting in both under-parameterized and overparameterized settings. To validate our theoretical analysis, we conducted simulation experiments on both linear regression models and Deep Neural Networks (DNNs). Results from these simulations substantiate our theoretical findings. Copyright 2024 by the author(s)

关键词： Multiple linear regression

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：