检索结果-内蒙古大学图书馆

Machine Learning for CAD (MLCAD), ACM/IEEE Workshop on

作者： Yishuang Lin Yaguang Li Meghna Madhusudan Sachin S. Sapatnekar Ramesh Harjani Jiang Hu Dept. of Computer Science and Engineering Texas A&M University Dept. of Electrical and Computer Engineering Texas A&M University Dept. of Electrical and Computer Engineering University of Minnesota

Performance modeling is a key bottleneck for analog design automation. Although machine learning-based models have advanced the state-of-the-art, they have so far suffered from huge data preparation cost, very limited reusability, and inadequate accuracy for large circuits. We introduce ML-based macro-modeling techniques to mitigate these problems for linear analog ICs and ADC/DACs. On representative testcases, our method achieves more than 1700× speedup for data preparation and remarkably smaller model errors compared to recent ML approaches. It also attains 3600× acceleration over SPICE simulation with very small errors and reduces data preparation time for an ADC design from 40 days to 9.6 hours.

关键词：

来源：评论

学校读者我要写书评

暂无评论

LipCraft: Building a Visual Speech Recognition System with Machine Learning and Streamlit

SSRN

引用

SSRN 2024年

作者： Karthikeyan, S. Sompalyam, Saivarshi Raveendran, Lekshmi S. Dept. of Computer Science and Engineering Sathyabama Institute of Science and Technology Tamil Nadu Chennai India

In the tapestry of human communication, speech has long been regarded as a cornerstone, seamlessly weaving the fabric of understanding and connection. Yet, for individuals facing hearing impairments or navigating environments where audible communication is limited, the reliance on traditional speech recognition technologies proves restrictive. Enter the realm of Visual Speech Recognition (VSR), an innovative domain seeking to decipher spoken language by interpreting the nuanced movements of the lips. In the crucible of this cutting-edge pursuit, our research takes centre stage with the project "LipCraft: Building a Visual Speech Recognition System with Machine Learning.". This project," LipCraft: Building a VSR System with Machine Learning and Streamlit Integration," stands at the convergence of these technologies, offering a holistic approach to redefine the boundaries of communication. The genesis of LipCraft is fuelled by a profound motivation rooted in inclusivity and technological innovation. Recognizing the constraints of traditional speech recognition, particularly for those with hearing impairments, the project aims to transcend these limitations by harnessing the power of deep learning, Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM) networks, and the unique capabilities of a Connectionist Temporal Classification (CTC) decoder. The scope of LipCraft extends beyond theoretical exploration into the practical realm, encompassing applications in accessibility services, education, and content creation. The project's innovative integration of technologies is not confined to the realms of machine learning and deep learning alone. LipCraft also introduces a user-friendly application interface developed with the Streamlit framework, demonstrating a commitment to making the technology accessible and user centric. LipCraft's objectives are ambitious and multifaceted. At its core objective to elevate the precision and efficiency of visual speech rec

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Leveraging compiler optimization for code clone detection 33

Leveraging compiler optimization for code clone detection

引用

33rd International Conference on Software engineering and Knowledge engineering, SEKE 2021

作者： Singh, Shirish Singhal, Harshit Mishra, Bharavi Dept. of Computer Science Columbia University United States Dept. of Computer Science and Engineering The LNMIIT India

ISBN: (纸本)1891706527

Finding similar code in software systems can guide several software engineering tasks such as code maintenance, program understanding, and code reuse. Similar code detection has been actively studied in the past. In the paper, we propose a novel approach that leverages compiler optimizations to transform semantically similar code and detect similar programs. The key observation of our work is that the compiler optimizations can be used to smooth out source code level idiosyncrasies introduced by the developers, thus making the optimized programs, for the same task, similar in structure. The similarity in structure can then be used to classify the programs. We conducted experiments on the Google CodeJam dataset to demonstrate the effectiveness of our approach. The experimental results show that our technique can achieve up to 85% accuracy on the program classification task, which is an improvement of more than 25% over the source code level classification. © 2021 Knowledge Systems Institute Graduate School. All rights reserved.

关键词： computer software reusability

来源：评论

学校读者我要写书评

暂无评论

Artificial Intelligence to Enhance the Security of Internet of Things: A Comprehensive Review

Artificial Intelligence to Enhance the Security of Internet ...

引用

Intelligent Data Communication Technologies and Internet of Things (IDCIoT), International Conference on

作者： Dheekshaa V M Harsha V Sindhu M S. Pavithra Dept. of Computer Science and Engineering Chennai Institute of Technology Chennai Tamil Nadu India

ISBN: (数字)9798331527549

ISBN: (纸本)9798331527556

The Internet of Things is the term we call when devices are connected to the network and work together to provide a better experience for users, facilitate better decision-making, and improve operations. As with any invention, this one also has its disadvantages. IoT devices have complex structures and low processing capacities. There are serious security risks in designing IoT like poor authentication, data leakage, and DDoS attacks. In this review paper, how AI can secure IoT is discussed. A few AI-driven strategies that help mask all these weaknesses are also discussed regarding intrusion detection systems (IDS), behavioural analytics, anomaly detection, and real-time threat monitoring. It presents a four-layered IoT architecture consisting of perception, network, support, and application layers. The use of the IoT systems will become more flexible and efficient in data security and collection, secure and encrypted communication, and predictive security analytics. AI will play the same role in improving the threat perception, real-time identification, and response, as this will make it indispensable to ensure the IoT environment security. This work emphasizes the urgent need for solid AI-powered, multilayered adaptive security frameworks to match the evolving demands of cybersecurity.

关键词： Technological innovation Reviews Decision making Solids Real-time systems Internet of Things Security Artificial intelligence Protection Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

A Secure and Efficient Group Authentication Framework for Blockchain-Enabled Vehicular Edge Computing Systems 3

A Secure and Efficient Group Authentication Framework for Bl...

引用

3rd International Conference on Disruptive Technologies, ICDT 2025

作者： Meena, Pawan Dash, Suryansh Sachin Monika Rani, Poonam Netaji Subhas University of Technology Dept. of Computer Science and Engineering New Delhi India

ISBN: (纸本)9798331519582

Maintaining privacy and trust within Vehicular Edge Computing (VEC) systems requires proper authentication methods that ensure safety and effectiveness in dynamic settings. Current authentication solutions face three main drawbacks including excessive computational costs together with complex identity management systems and large security exposure. The proposed lightweight Multi-Entity Authenticated Group Key Agreement protocol serves blockchain-based VEC networks according to this research. MAGKA enhances authentication through bulk certification removal and reduces reliance on identity authentication and provides seamless vehicle traffic across edge networks. The solution provides VEC authentication with robust security and scalability demonstrated through security analysis of its ability to withstand attacks. © 2025 IEEE.

关键词： Ad hoc networks

来源：评论

学校读者我要写书评

暂无评论

A 10-Bit Successive Approximation Register Analog to Digital Converter Using VCO-Based Time-Domain Comparator

A 10-Bit Successive Approximation Register Analog to Digital...

引用

Consumer Technology (ICCT-Pacific), International Conference on

作者： Ko-Chi Kuo Jia-Wei Chen Dept. Computer Science and Engineering National Sun Yat-sen University Kaohsiung Taiwan

ISBN: (数字)9798331504120

ISBN: (纸本)9798331504137

The successive approximation register (SAR) analog-to-digital converter (ADC) architecture is employed to achieve low power consumption, allowing it to operate at low supply voltage. A time-domain quantization architecture based on a time-domain comparator is combined to overcome the design bottlenecks caused by the reduced intrinsic gain of transistors, the increased process variation, and the decreased supply voltage in advanced processes. A voltage-to-time converter based on a voltage-controlled oscillator (VCO) is used, providing the benefits of circuit stability and low noise impact. This design can be further improved by adding an SR time amplifier to the output of VCO, allowing the phase difference to grow exponentially. Hence, a faster escape from the dead zone of the phase detector can be realized to resolve the issue between high resolution and high speed in traditional VCO comparators. To accommodate the output generated by the improved voltage-to-time converter, the operation mode of the phase detector is modified, enabling it to be applied to the new voltage-to-time converter output. This paper presents a 10-bit SAR ADC in TSMC 40nm process, with a conversion rate of 10MS/s, a supply voltage of 0.9V, power consumption of 57.9µW, an ENOB of 9.88 bits, an SFDR of 74.07dB, and a FoM of 6.14 fJ/conv.-step.

关键词： Quantization (signal) Power demand Voltage-controlled oscillators Noise Detectors Voltage Registers Circuit stability Time-domain analysis Analog-digital conversion

来源：评论

学校读者我要写书评

暂无评论

Duality between Lines and Points

arXiv

引用

arXiv 2025年

作者： Saxena, Sanjeev Dept. of Computer Science and Engineering Indian Institute of Technology Kanpur208 016 India

There are several notions of duality between lines and points. In this note, it is shown that all these can be studied in a unified way. Most interesting properties are independent of specific choices. It is also shown that either dual mapping can be its own inverse or it can preserve relative order (but not both). Generalisation to higher dimensions is also discussed. An elementary and very intuitive treatment of relationship between arrangements in d + 1 dimensions and searching for k-nearest neighbour in d-dimensions is also given. Copyright © 2025, The Authors. All rights reserved.

关键词： Nearest neighbor search

来源：评论

学校读者我要写书评

暂无评论

DMDA: A Computational Resource Allocation Approach for IoT Devices in Fog Computing

DMDA: A Computational Resource Allocation Approach for IoT D...

引用

Intelligent Cybernetics Technology & Applications (ICICyTA), International Conference on

作者： Luthfan Hadi Pramono Shan-Hsiang Shen Dept. of Computer Science and Information Engineering National Taiwan University of Science and Technology Taipei Taiwan

ISBN: (数字)9798331506490

ISBN: (纸本)9798331506506

Fog Computing enables the efficient offloading of computational tasks from IoT Devices to Fog nodes, enhancing processing efficiency and reducing latency. This paper introduces the Dynamic Matching with Deferred Acceptance (DMDA) algorithm, designed to optimize the allocation of GPU and CPU resources within Fog Computing environments, specifically for smart city applications. The DMDA algorithm allocates resources to IoT Devices based on parameters such as data type, computational load, priority, and responsiveness. Devices processing image or video data are given priority for GPU allocation, while other devices are assigned to CPUs based on their cumulative performance scores. Devices with lower priority are reassigned to CPU resources or excluded when resource availability is exhausted, significantly enhancing performance under high-demand scenarios. The algorithm employs a Mixed-Integer Linear Programming (MILP) model to maximize the aggregate score of allocated devices while adhering to resource constraints. Experiments conducted with 50 to 1,000 devices demonstrate that both the total Device Score rate and Allocated Device rate reach 100%, indicating optimal allocation. Empirical findings show that the DMDA algorithm improves resource utilization, reduces processing times, and ensures effective workload distribution across large-scale IoT deployments.

关键词： Performance evaluation Smart cities Heuristic algorithms Graphics processing units Parallel processing Dynamic scheduling Real-time systems Resource management Internet of Things Edge computing

来源：评论

学校读者我要写书评

暂无评论

Exploring Large Language Models for Semantic Analysis and Categorization of Android Malware

arXiv

引用

arXiv 2025年

作者： Walton, Brandon J. Khatun, Eshita Ghawaly, James M. Ali-Gombe, Aisha Dept. of Computer Science & Engineering Louisiana State University Baton Rouge United States

Malware analysis is a complex process of examining and evaluating malicious software’s functionality, origin, and potential impact. This arduous process typically involves dissecting the software to understand its components, infection vector, propagation mechanism, and payload. Over the years, deep reverse engineering of malware has become increasingly tedious, mainly due to modern malicious codebases’ fast evolution and sophistication. Essentially, analysts are tasked with identifying the elusive needle in the haystack within the complexities of zero-day malware, all while under tight time constraints. Thus, in this paper, we explore leveraging Large Language Models (LLMs) for semantic malware analysis to expedite the analysis of known and novel samples. Built on GPT-4o-mini model, MalParse is designed to augment malware analysis for Android through a hierarchical-tiered summarization chain and strategic prompt engineering. Additionally, MalParse performs malware categorization, distinguishing potential malware from benign applications, thereby saving time during the malware reverse engineering process. Despite not being fine-tuned for Android malware analysis, we demonstrate that through optimized and advanced prompt engineering MalParse can achieve up to 77% classification accuracy while providing highly robust summaries at functional, class, and package levels. In addition, leveraging the backward tracing of the summaries from package to function levels allowed us to pinpoint the precise code snippets responsible for malicious behavior. Copyright © 2025, The Authors. All rights reserved.

关键词： Android malware

来源：评论

学校读者我要写书评

暂无评论

Facial Forgery-Based Deepfake Detection Using Fine-Grained Features 22

Facial Forgery-Based Deepfake Detection Using Fine-Grained F...

引用

22nd IEEE International Conference on Machine Learning and Applications, ICMLA 2023

作者： Nadimpalli, Aakash Varma Rattani, Ajita School of Computing Wichita State University Kansas United States Uni. of North Texas at Denton Dept. of Computer Science and Engineering TX United States

ISBN: (纸本)9798350345346

Facial forgery by deepfakes has caused major se-curity risks and raised severe societal concerns. As a counter-measure, a number of deepfake detection methods have been proposed. Most of them model deepfake detection as a binary classification problem using a backbone convolutional neural network (CNN) architecture pretrained for the task. These CNN-based methods have demonstrated very high efficacy in deepfake detection with the Area under the Curve (AUC) as high as 0.99. However, the performance of these methods degrades signifi-cantly when evaluated across datasets and deepfake manipulation techniques. This draws our attention towards learning more subtle, local, and discriminative features for deepfake detection. In this paper, we formulate deepfake detection as a fine-grained classification problem and propose a new fine-grained solution to it. Specifically, our method is based on learning subtle and generalizable features by effectively suppressing background noise and learning discriminative features at various scales for deepfake detection. Through extensive experimental validation, we demonstrate the superiority of our method over the published research in cross-dataset and cross-manipulation generalization of deepfake detectors for the majority of the experimental scenarios. © 2023 IEEE.

关键词： Cross-Manipulation Generalization Deepfakes Facial Manipulations Fine-Grained Classification

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：