检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Franco, Nicola Korth, Daniel Lorenz, Jeanette Miriam Roscher, Karsten Günnemann, Stephan Fraunhofer Institute for Cognitive Systems IKS Munich Germany Dept. of Computer Science Munich Data Science Institute Technical Univ. of Munich Germany

As the use of machine learning continues to expand, the importance of ensuring its safety cannot be overstated. A key concern in this regard is the ability to identify whether a given sample is from the training distribution, or is an "Out-Of-Distribution" (OOD) sample. In addition, adversaries can manipulate OOD samples in ways that lead a classifier to make a confident prediction. In this study, we present a novel approach for certifying the robustness of OOD detection within a 2-norm around the input, regardless of network architecture and without the need for specific components or additional training. Further, we improve current techniques for detecting adversarial attacks on OOD samples, while providing high levels of certified and adversarial robustness on in-distribution samples. The average of all OOD detection metrics on CIFAR10/100 shows an increase of ∼ 13%/5% relative to previous approaches. Code: https://***/FraunhoferIKS/distro © 2023, CC BY.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Smart Building Indoor Temperature Prediction Using the IoT and Machine Learning 2

Smart Building Indoor Temperature Prediction Using the IoT a...

引用

2nd International Conference on Electrical, Electronics, Information and Communication Technologies, ICEEICT 2023

作者： Vincent, Robin Rohit Sarkar, Swagata Irence, Shalom Prakash, R.B.R. Senthil Rama, R. Suganthi, D. School of Computer Science & Engineering and Information Science Computer Science & Engineering Presidency University Bengaluru560064 India Artificial Intelligence and Data Science Sri Sairam Engineering College Chennai 44 India EEE Karunya Kits / CBE India Koneru Lakshmaiah Education Foundation Electrical and Electronics Engineering Vaddeswaram522302 India DMI College of Engineering EEE Chennai India Saveetha College of Liberal Arts and Sciences Department of Computer Science Saveetha Institute of Medical and Technical Sciences Chennai India

ISBN: (纸本)9798350397635

Due to fast expansion and improved lifestyle, the need for useable energy has skyrocketed in recent decades, notably in the construction industry. Several factors, such as local climate, building makeup, and energy consumption habits, affect a structure's efficiency in using energy. In this paper, we aim for forecasting the interior temperature of smart buildings by combining IoT with popular machine Learning algorithms. This prediction model has been developed using online learning techniques to increase its adaptability to novel inputs. This paper details a methodology for creating an analytical model that may be utilized for smart building interior temperature forecasting by merging the IoT with some popular machine learning methods. This prediction model was built with an online-learning strategy, so it may be used for a dataset with which the developers are unfamiliar. To ensure the precision of the method, this article uses Machine Learning to test it using real-world sensor data. In order to verify the methodology, the study conducts experiments using Machine Learning on collected, real-world sensor data. The study then proposes incorporating the following method into an Edge Computing based IoT architecture in order to make the building operational in an energy-efficient manner. The learning framework was validated with data collected from an Internet of Things network installed in a home over the course of a summertime. The learnt model's efficacy has been evaluated, and the model's predictability of interior temperature changes has been quantified. The paper demonstrates the viability of the proposed Learning framework for thermal models based on the Internet of Things and provides a pragmatic answer to the problem of supplying accurate thermal models for use in the temperature controls of smart buildings of the future. © 2023 IEEE.

关键词： Energy utilization

来源：评论

学校读者我要写书评

暂无评论

Efficient MILP Decomposition in Quantum Computing for ReLU Network Robustness

Efficient MILP Decomposition in Quantum Computing for ReLU N...

引用

Quantum Computing and Engineering (QCE), IEEE International Conference on

作者： Nicola Franco Tom Wollschläger Benedikt Poggel Stephan Günnemann Jeanette Miriam Lorenz Fraunhofer Institute for Cognitive Systems IKS Munich Germany Dept. of Computer Science & Munich Data Science Institute Technical Univ. of Munich Germany

Emerging quantum computing technologies, such as Noisy Intermediate-Scale Quantum (NISQ) devices, offer potential advancements in solving mathematical optimization problems. However, limitations in qubit availability, noise, and errors pose challenges for practical implementation. In this study, we examine two decomposition methods for Mixed-Integer Linear Programming (MILP) designed to reduce the original problem size and utilize available NISQ devices more efficiently. We concentrate on breaking down the original problem into smaller subproblems, which are then solved iteratively using a combined quantum-classical hardware approach. We conduct a detailed analysis for the decomposition of MILP with Benders and Dantzig-Wolfe methods. In our analysis, we show that the number of qubits required to solve Benders is exponentially large in the worst-case, while remains constant for Dantzig-Wolfe. Additionally, we leverage Dantzig-Wolfe decomposition on the use-case of certifying the robustness of ReLU networks. Our experimental results demonstrate that this approach can save up to 90% of qubits compared to existing methods on quantum annealing and gate-based quantum computers.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Advanced Deep Learning Model for Enhanced Kidney Disease Detection 5

Advanced Deep Learning Model for Enhanced Kidney Disease Det...

引用

5th International Conference on data Intelligence and Cognitive Informatics, ICDICI 2024

作者： Davamani, K. Anita Dhanya, D. Pushparani, S. Vidhya, R. Mohan, Geetha Jani Anbarasi, L. Saveetha Institute of Medical and Technical Sciences Saveetha School of Engineering Chennai India Mar Ephraem College of Engineering and Technology Department of Artificial Intelligence and Data Science Elavuvilai India Meenakshi College of Engineering Department of Cse Chennai India Agni College of Technology Department of Information Technology Chennai India Institute of Applied Technology Abu Dhabi United Arab Emirates Vellore Institute of Technology School of Computer Science and Engineering Chennai India

ISBN: (纸本)9798350389609

Automated Chronic Kidney Disease (CKD) is a swiftly growing non-communicable condition that contributes to a substantial number of deaths and illnesses worldwide. Affecting over 10% of the global population, CKD results in millions of fatalities each year. In this study, an advanced deep learning model for kidney disease detection is proposed that combines convolutional and involution layers to improve classification accuracy. The model processes input images by first applying a 2D convolution with 32 filters of size 3x3, followed by ReLU activation to extract key features. Max-pooling layers are then used to reduce spatial dimensions while retaining crucial information. The convolutional layers are sequentially stacked with larger filter sizes, resulting in progressively richer feature representations. To enhance feature extraction, the model integrates custom involution layers, which extend traditional convolution operations and are particularly effective for processing large-scale image data. These involution layers are applied after the convolutional steps, utilizing 128 and 256 filters of size 3 × 3, followed by max-pooling to further refine the extracted features. A comparative analysis was carried out between the proposed custom Convolutional Neural Network (CNN) model and previous research works, utilizing metrics such as Recall, Precision, F1-Score, Accuracy, Loss, Area Under the Curve (AUC) and Receiver Operating Characteristic (ROC) curves. The results indicate that the proposed model achieves superior classification prediction performance, with an accuracy of 99% and reduced computational complexity. Additionally, it demonstrates a enhanced F1-Score of 100%, 98%, 98% and 100% for Cyst, Normal, Stone and Tumor respectively. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

TailClip: Mitigating Tail Latency in Cloud Gaming via Smart Video Frame Generation

TailClip: Mitigating Tail Latency in Cloud Gaming via Smart ...

引用

IEEE International Conference on Multimedia and Expo (ICME)

作者： Qianyun Gong Kunheng Jiang Jingjing Wen Xinjing Yuan Jianxin Shi Lingjun Pu Institute of Systems and Networks College of Computer Science Nankai University Tianjin China Key Laboratory of Data and Intelligent System Security Ministry of Education Nankai University Tianjin China

ISBN: (数字)9798350390155

ISBN: (纸本)9798350390162

Latency is one of the most significant issues in cloud gaming, among which tail latency, mainly attributed to dynamic network environments (i.e., transmission) and limited device computing capacity (e.g., decoding), has attracted increasing attention. To mitigate the tail latency, different from existing researches considering resource adaptation such as bitrate adaptation, we propose TailClip, whose novel idea is to enable video frame generation at the client if the tail latency is about to happen. TailClip consists of two innovative components: a Deep Reinforcement Learning (DRL) driven tail latency trigger that jointly decides a series of subsequent frame generation regarding multidimensional features of historical tail latency; a lightweight frame generation model derived by adaptive pruning in terms of device computing capacity at runtime. Extensive evaluations indicate the superior performance of TailClip. For example, it can remove the high-latency frames (i.e., over 100 ms) by 77% with an acceptable video quality (i.e., 0.45 dB reduction on average).

关键词： Performance evaluation Runtime Computational modeling Tail Cloud gaming Streaming media Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Monte Carlo Tree Search Algorithms for Strategic Planning in Humanoid Robotics 1

Monte Carlo Tree Search Algorithms for Strategic Planning in...

引用

1st IEEE International Conference on Cognitive Robotics and Intelligent Systems, ICC - ROBINS 2024

作者： Gundla, Praveen Gummadi, Annapurna Changala, Ravindra Vuyyuru, VeeraAnkalu Hemantha Kumar, A. Kiranbala, B. Nri Institute of Technology Andhra Pradesh Guntur India Hyderabad India Guru Nanak Institutions Technical Campus Department of Information Technology Hyderabad India KoneruLakshmaiah Education Foundation Department of Computer Science and Engineering A.P Vaddeswaram522502 India MED Annamacharya Institute of Technology and Sciences AP Rajampet India K.Ramakrishnan College of Engineering Department of Artificial Intelligence and Data Science Trichy India

ISBN: (纸本)9798350372748

With the increasing complexity and versatility of humanoid robots, there is a growing imperative for efficient and adaptable planning methodologies. This evolution marks a crucial step towards enhancing the overall capabilities and practicality of humanoid robotic systems. This research delves into the application of Monte Carlo Tree Search (MCTS) algorithms for enhancing strategic planning in the context of humanoid robotics. The study explores the integration of MCTS into humanoid robotics to enable effective decision-making processes that account for uncertainties, sensor noise, and real-time adaptability. By simulating future scenarios and evaluating potential actions through iterative sampling, MCTS facilitates the creation of robust and context-aware strategies. The research considers various aspects of strategic planning, including path finding, task allocation, and goal achievement. This investigation contributes to the burgeoning field of robotics by providing a comprehensive analysis of the benefits and challenges associated with implementing MCTS in humanoid robotic systems. The proposed MCTSA model shows better efficiency in Resource Allocation 88%, SuccessRate (Navigation) 85%, Average Planning Time 2.5 seconds and Success Rate (Object Manipulation) 92% which is higher when compared to the existing methods. The findings aim to inform the development of intelligent and autonomous robots capable of making strategic decisions in dynamic and uncertain environments, thereby advancing the capabilities of humanoid robotics in various practical applications. © 2024 IEEE.

关键词： Decision making

来源：评论

学校读者我要写书评

暂无评论

Research on the Method of Constructing Knowledge Graph of Affective Disorder Based on Multi-Source data Fusion 5

Research on the Method of Constructing Knowledge Graph of Af...

引用

5th International Conference on Information Technologies and Electrical Engineering, ICITEE 2022

作者： Wang, Yang Yang, Weihong School of Artificial Intelligence and Big Data Luzhou Vocational and Technical College Sichuan Luzhou646000 China School of Computer Science and Technology Harbin Institute of Technology Heilongjiang Harbin150006 China

ISBN: (纸本)9781450396806

Affective disorder (Depression) has the characteristics of high incidence rate, low recognition rate, high disability and death rate, and has become a serious public health problem that needs to be solved urgently. It is particularly important to use machine learning and big data mining technology to build a complete knowledge graph of affective disorders. In this paper, multi-source data fusion is used to construct the original corpus of the knowledge graph of affective disorders. In the information extraction stage, we designed a machine learning algorithm based on Hidden Markov Model (HMM). The algorithm can extract entities and entity mapping relationships from corpus to form entity database and triplet relationship database. In the knowledge fusion stage, multi-source heterogeneous knowledge fusion algorithm is used to complete entity alignment and relationship deduction. Finally, the open-source graphic database Neo4j is used as the underlying storage structure, so as to realize the visualization of the knowledge graph of affective disorders. © 2022 ACM.

关键词： data fusion

来源：评论

学校读者我要写书评

暂无评论

Randomized Message-Interception Smoothing: Gray-box Certificates for Graph Neural networks

arXiv

引用

arXiv 2023年

作者： Scholten, Yan Schuchardt, Jan Geisler, Simon Bojchevski, Aleksandar Günnemann, Stephan Dept. of Computer Science Munich Data Science Institute Technical University of Munich Germany CISPA Helmholtz Center for Information Security Germany

Randomized smoothing is one of the most promising frameworks for certifying the adversarial robustness of machine learning models, including Graph Neural networks (GNNs). Yet, existing randomized smoothing certificates for GNNs are overly pessimistic since they treat the model as a black box, ignoring the underlying architecture. To remedy this, we propose novel gray-box certificates that exploit the message-passing principle of GNNs: We randomly intercept messages and carefully analyze the probability that messages from adversarially controlled nodes reach their target nodes. Compared to existing certificates, we certify robustness to much stronger adversaries that control entire nodes in the graph and can arbitrarily manipulate node features. Our certificates provide stronger guarantees for attacks at larger distances, as messages from farther-away nodes are more likely to get intercepted. We demonstrate the effectiveness of our method on various models and datasets. Since our gray-box certificates consider the underlying graph structure, we can significantly improve certifiable robustness by applying graph sparsification.1 © 2023, CC BY-SA.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Wavelength-division-multiplexed Transmission and Dispersion Compensation Enabled by Soliton Microcombs and Cascaded Microrings

Wavelength-division-multiplexed Transmission and Dispersion ...

引用

2024 Conference on Lasers and Electro-Optics, CLEO 2024

作者： Liu, Yuanbin Zhang, Hongyi Liu, Jiacheng Lu, Liangjun Du, Jiangbing Li, Yu He, Zuyuan Chen, Jianping Zhou, Linjie Poon, Andrew W. State Key Laboratory of Advanced Optical Communication Systems and Networks Shanghai Key Lab of Navigation and Location Services Shanghai Institute for Advanced Communication and Data Science Department of Electronic Engineering Shanghai Jiao Tong University Shanghai200240 China SJTU-Pinghu Institute of Intelligent Optoelectronics Pinghu314200 China Photonic Device Laboratory Department of Electronic and Computer Engineering The Hong Kong University of Science and Technology Clear Water Bay Hong Kong

ISBN: (纸本)9781957171395

We demonstrate wavelength-division-multiplexed data transmission and dispersion compensation of 25 Gb/s × 9 on-off-keying signals over a 20-km singlemode fiber using an integrated single-soliton microcomb and a cascaded microring resonator-based reconfigurable dispersion compensator. © Optica Publishing Group 2024, © 2024 The Author(s)

关键词： Optical fibers

来源：评论

学校读者我要写书评

暂无评论

Wavelength-division-multiplexed Transmission and Dispersion Compensation Enabled by Soliton Microcombs and Cascaded Microrings

Wavelength-division-multiplexed Transmission and Dispersion ...

引用

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

关键词： Dispersion compensation

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：