检索结果-内蒙古大学图书馆

computer Security Applications Conference Workshops (ACSAC Workshops), Annual

作者： Bikash Saha Nanda Rani Sandeep Kumar Shukla Dept. of Computer Science and Engineering Indian Institute of Technology Kanpur Kanpur India

ISBN: (数字)9798331532819

ISBN: (纸本)9798331532826

The increasing reliance on Industrial Internet of Things (IIoT) systems in critical infrastructure has made these environments prime targets for stealthy and prolonged Advanced Persistent Threats (APTs) that evade traditional security systems. This study introduces MAD (Meta-learning based APT Detection), a novel framework designed to detect APTs in IIoT environments by leveraging provenance data. This data provides a detailed view of system-level interactions, capturing subtle and persistent malicious behaviors. We evaluate MAD on the CICAPT-IIoT dataset, which simulates real-world APT scenarios with severe class imbalance due to the "low and slow" nature of APT attacks. MAD includes two variants: MAD v1, utilizing a multi-layer perceptron (MLP), and MAD v2, employing Model-Agnostic Meta-Learning (MAML) to improve detection of rare attack behaviors. MAD v1 achieves F1-score of 0.91 and outperforms the existing baselines. MAD v2 achieves F1-score of 0.95 surpassess existing baselines as well as MAD v1, which demonstrates its enhanced capability to detect rare APT activities despite the dominance of benign samples. This work combines provenance-based detection with meta-learning to address class imbalance and enhance APT detection in IIoT, demonstrating MAD’s potential in tackling emerging IIoT security challenges.

关键词： Metalearning Deep learning Adaptation models Adaptive systems Conferences Threat assessment Real-time systems Critical infrastructure computer security Industrial Internet of Things

来源：评论

学校读者我要写书评

暂无评论

引用

International Symposium on Computing and Networking Workshops (CANDARW)

作者： Yashdeep Singh Sushanta Karmakar Dept. of Computer Science and Engineering Indian Institute of Technology Guwahati Guwahati India

ISBN: (数字)9798331505349

ISBN: (纸本)9798331505356

We solve the following problem: given two sets of different items and a set of people with combined preferences over pairs of items, one from each set, what is the best way to match the pair of items to the people? To solve this problem, we consider the popular matching problem in a 3-uniform 3-partite hypergraph, where the first set comprises agents, while the second and third sets contain distinct items respectively. Agents have preferences over the hyperedges, whereas the members of the other sets lack such preferences. An agent votes for a matching in which he/she obtains a better assignment. A matching M is called popular if there is no matching M' such that the number of agents that prefer M' is more than the number of agents that prefer M. Finding a popular matching, if it exists, is a hard problem. Therefore, one approach is to find a matching that is approximately popular. The usual metric in this case is called the unpopularity factor of a matching M, denoted by u(M). In this paper, we provide an approximation algorithm that produces a matching M in a 3-uniform 3-partite hypergraph such that $u(M) \leq 1.5 \times {3^{\frac{k}{2}}}$ where k is the maximum degree of an agent. The approximation algorithm takes O(k(n 1.5 + m)) time to produce the matching M where n is the total number of agents and m is the total number of hyperedges. In a 3-uniform 3-partite hypergraph where k is a constant, our algorithm performs well and produces a matching M with u(M) = O(1). Importantly, our approach stands as the inaugural algorithm to achieve a bounded unpopularity matching in a hypergraph.

关键词： Measurement Conferences Approximation algorithms

来源：评论

学校读者我要写书评

暂无评论

Compressive Recovery of Signals Defined on Perturbed Graphs

arXiv

引用

arXiv 2024年

作者： Ghosh, Sabyasachi Rajwade, Ajit Dept. of Computer Science and Engineering Indian Institute of Technology Bombay Mumbai India

Recovery of signals with elements defined on the nodes of a graph, from compressive measurements is an important problem, which can arise in various domains such as sensor networks, image reconstruction and group testing. In some scenarios, the graph may not be accurately known, and there may exist a few edge additions or deletions relative to a ground truth graph. Such perturbations, even if small in number, significantly affect the Graph Fourier Transform (GFT). This impedes recovery of signals which may have sparse representations in the GFT bases of the ground truth graph. We present an algorithm which simultaneously recovers the signal from the compressive measurements and also corrects the graph perturbations. We analyze some important theoretical properties of the algorithm. Our approach to correction for graph perturbations is based on model selection techniques such as cross-validation in compressed sensing. We validate our algorithm on signals which have a sparse representation in the GFT bases of many commonly used graphs in the network science literature. An application to compressive image reconstruction is also presented, where graph perturbations are modeled as undesirable graph edges linking pixels with significant intensity difference. In all experiments, our algorithm clearly outperforms baseline techniques which either ignore the perturbations or use first order approximations to the perturbations in the GFT bases. Copyright © 2024, The Authors. All rights reserved.

关键词： Recovery

来源：评论

学校读者我要写书评

暂无评论

Opportunistic Migration for Hybrid Memories While Mitigating Aging Effects

Opportunistic Migration for Hybrid Memories While Mitigating...

引用

IEEE International Conference on computer Design: VLSI in computers and Processors, (ICCD)

作者： Aswathy N S Hemangee K. Kapoor Dept. of Computer Science and Engineering Indian Institute of Technology Guwahati Guwahati Assam

ISBN: (数字)9798350380408

ISBN: (纸本)9798350380415

Hybrid memory systems composed of Non-volatile memory (NVM) and DRAM to exploit the high density of NVM and low access latency of DRAM. Phase Change Memory (PCM), a type of non-volatile memory, is a viable choice for main memory. High write latency and high voltage requirements for PCM lead to Biased Temperature Instability (BTI) aging and performance degradation. De-stressing the memory circuit at regular intervals controls BTI aging. Memory performance can be enhanced by migrating the highest write count memory pages across memory units. Migration and de-stress halt the service of regular requests and affect the performance of the system. Therefore, it is crucial to control migration and de-stress to enhance hybrid memory performance while mitigating BTI aging. We propose DOPMig, a de-stress-aware page migration technique. The policy migrates write-intensive pages to DRAM at regular intervals but opportunistically parallel to the de-stress operation. This method of background migration helps to reduce the migration overheads and improves performance. DOPMig achieves a performance gain of 22%, and improves memory service rate by 15%, and increases DRAM access by 24%.

关键词： Phase change materials Degradation Nonvolatile memory Memory management Random access memory High-voltage techniques Aging Performance gain Phase change memory Proposals

来源：评论

学校读者我要写书评

暂无评论

Detection of Geometry-based Adversarial Facial Attacks using Error Level Analysis

Detection of Geometry-based Adversarial Facial Attacks using...

引用

2022 IEEE Region 10 International Conference, TENCON 2022

作者： Sadu, Chiranjeevi Das, Pradip K. Rajiv Gandhi University of Knowledge Technologies Dept. of Computer Science and Engineering Nuzvid India Indian Institute of Technology Guwahati Dept. of Computer Science and Engineering Guwahati India

ISBN: (数字)9781665450959

ISBN: (纸本)9781665450959

Despite Machine Learning (ML) and Deep Learning (DL) models in specific have achieved significant success in various applications of computer vision in recent years, they remain vulnerable to carefully crafted, human-invisible adversarial attacks. Adversarial attacks are generated intentionally by adding small perturbations to an input image to mislead the classifier. This paper investigates a new defense method against geometry-based adversarial face attacks that significantly mitigate the impact of such adversarial attacks on image-classification models. We generate geometry-based adversarial face images based on Fast Landmark Manipulation (FLM) method and Error Level Analysis (ELA) is performed to effectively discriminate adversarial face images from clean images. The performance of the proposed defense method is demonstrated on a real-world dataset with different types of classifiers. The experimental results show that our defense method can effectively detect adversarial face images from clean images with 99.75% accuracy. © 2022 IEEE.

关键词： Geometry

来源：评论

学校读者我要写书评

暂无评论

Improving Time Series Forecasting Accuracy with Transformers: A Comprehensive Analysis with Explainability

Improving Time Series Forecasting Accuracy with Transformers...

引用

Electrical, Electronics, Information and Communication Technologies (ICEEICT), International Conference on

作者： Susmera Joseph Ashly Ann Jo Ebin Deni Raj Dept. of Computer Science Engineering Indian Institute of Information Technology Kottayam India

ISBN: (数字)9798350369083

ISBN: (纸本)9798350369090

Time series forecasting is crucial in numerous sectors, including healthcare, energy, and finance. Transformer models, initially designed for natural language processing applications, demonstrate potential in capturing temporal dependencies within sequential data. In this paper, the application of transformer models in time-series forecasting is explored, and their performance and interpretability are investigated. Transparent models build trust and facilitate decision-making, especially in industries with ethical guidelines. While inherently interpretable time-series Transformers exist, there is a gap in research regarding post hoc methods, which is an area ripe for exploration. Understanding and comparing complex Transformer models for time series forecasting can benefit from advancements in explainability, ensuring accurate and comprehensible predictions. In comparison to five other transformer-based models and classical models, the Temporal Fusion Transformer (TFT) model outperformed them in short-term forecasting. Further enhancement of the model performance was achieved through a meticulous analysis of the Encoder-Decoder’s importance, which contributed significantly to refining the forecasting accuracy. The vanilla transformer was explained using permutation feature importance. The model stability was analyzed using perturbation methods.

关键词： Analytical models Accuracy Perturbation methods Time series analysis Refining Predictive models Transformers Stability analysis Thin film transistors Forecasting

来源：评论

学校读者我要写书评

暂无评论

A Comprehensive Review of Routing Protocols in Vehicular Ad Hoc Networks (VANETs): Challenges, Solutions and Future Directions

A Comprehensive Review of Routing Protocols in Vehicular Ad ...

引用

IEEE UP Section Conference on Electrical computer and Electronics (UPCON)

作者： Sonika Bhardwaj Ramesh Saha Dept. of Computer Science & Engineering Indian Institute of Information Technology Sonepat India

ISBN: (数字)9798350378726

ISBN: (纸本)9798350378733

VANETs play a pivotal role in the growth/development of the intelligent transportation systems (ITS) by allowing seamless communication between vehicles and infrastructure. This paper presents a comprehensive review of various routing protocols used in VANETs, focusing on topology-based, position-based, hybrid, and machine learning (ML) enhanced protocols. The analysis evaluates these protocols across key performance metrics such as scalability, latency, packet delivery ratio (PDR), overhead, and security. Hybrid and context-aware protocols provide greater adaptability but come with more complexity than standard protocols, which struggle with scalability and security in dynamic contexts. The growing combination of ML, blockchain, and federated learning into VANET routing establishes potential for enhanced performance and security. Furthermore, the paper also highlights the current limitations of an existing protocols and recognizes the directions for the research, the more emphasizing on the need for scalable, secure, and real-time adaptable solutions for next-generation VANETs.

关键词： Federated learning Reviews Scalability Vehicular ad hoc networks Routing Routing protocols Real-time systems Blockchains Security Vehicle dynamics

来源：评论

学校读者我要写书评

暂无评论

A Detection Method for Copy-Move Forgery Attacks in Digital Images

A Detection Method for Copy-Move Forgery Attacks in Digital ...

引用

2022 IEEE Region 10 International Conference, TENCON 2022

ISBN: (数字)9781665450959

ISBN: (纸本)9781665450959

Digital images are carrying important information in many real-world applications such as surveillance, courts of law as evidence for crimes, journalism, scientific publications, and medical imaging. Manipulating these images becomes easy due to the paramount image processing and editing tools. Thus, the authentication of digital images becomes one of the principal problems in digital image forensics and attracts the attention of the research community. In this paper, we focus on the detection of copy-move forgery attacks in digital images. It is one of the most actively used attacks. It copies some region of the image and pastes it into one or more parts of the same image to cover or emphasize important scenes in the image. We propose a method to detect such forgery regions based on Speeded Up Robust Features (SURF) and Binary Robust Invariant Scalable Keypoints (BRISK) descriptors. Both fused features are matched and performed clustering to reduce the false positives using Hierarchical Agglomerative Clustering (HAC). The proposed method is tested on real-world copy-move datasets. Experimental results show that our method is robust against various geometric transformations and able to precisely determine the forged regions. © 2022 IEEE.

关键词： Medical imaging

来源：评论

学校读者我要写书评

暂无评论

Fault Testing in AI-Accelerators: A Review 33

Fault Testing in AI-Accelerators: A Review

引用

33rd IEEE Asian Test Symposium, ATS 2024

作者： Bhattacharya, Bhargab B. Das, Debesh K. Chatterjee, Subhajit Rahaman, Hafizur Indian Statistical Institute Acm Unit Kolkata700 108 India Jadavpur University India Computer Science & Engg. Dept. Kolkata700032 India Indian Institute of Engineering Sc. and Tech. School of Vlsi Technology Howrah711 103 India

ISBN: (纸本)9798331529161

With the emergence of all-inclusive AI/ML applications, hardware solutions, commonly known as AI-Accelerators (AIA), are now being widely adopted to emulate deep neural networks (DNN) to facilitate faster and large-scale data analytics. An AIA-chip comprises a 2D systolic array of identical processing units (PEs), registers, and glue logic. These arrays may be implemented with traditional digital logic or with analog primitives such as memristors. As the packing density of AIA-chips increases, they become vulnerable to various manufacturing defects thereby compromising yield and the accuracy of prediction. In this review article, we summarize various methods that have been recently proposed for expediting Automatic Test-Pattern Generation (ATPG) for stuck-at and transition faults in AIA-arrays. Other relevant issues such as fault-criticality, self-test, fault-recovery, and the asymmetry of fault behavior, are also discussed. © 2024 IEEE.

关键词： Systolic arrays

来源：评论

学校读者我要写书评

暂无评论

Advancements in Flood Mapping Methodologies: A Comprehensive Review

Advancements in Flood Mapping Methodologies: A Comprehensive...

引用

IEEE Recent Advances in Intelligent Computational Systems (RAICS)

作者： Divya V Chandran J Anitha Amrutha Vijayakumar Anjana Raveendran Dept. of Electronics and Communication Engineering Adi Shankara Institute of Engineering and Technology Kalady India Dept.of Computer Science Engineering Karunya Institute of Technology and Sciences Coimbatore India

ISBN: (数字)9798350381689

ISBN: (纸本)9798350381696

This review aims to comprehensively assess the recent advancements in flood mapping methodologies and technologies. By analyzing diverse techniques and approaches employed in flood mapping, this review seeks to present an in-depth overview of the current landscape of this field. The primary focus is on evaluating the strengths and limitations inherent in various flood mapping methods. Special attention will be given to exploring the integration of remote sensing technologies and geographic information systems (GIS) in flood mapping. Additionally, the review will delve into the transformative impact of machine learning and artificial intelligence on enhancing the precision and efficiency of flood mapping processes. Recent developments in flood mapping have significantly elevated the accuracy and reliability of the techniques employed. The integration of AI technologies with remote sensing methods has notably contributed to this advancement. Conventional approaches to flood mapping have traditionally relied on multiple methodologies to comprehend potential flood impacts. However, these methods often prove laborious, resource-intensive, and may lack access to real-time data, posing limitations for precise flood mapping and prediction. These identified limitations underscore the critical need for the development of more robust flood mapping techniques. Addressing these challenges is essential for enhancing flood risk assessment, improving management strategies, and ensuring better preparedness for future flood events.

关键词： Accuracy Reviews Heuristic algorithms Real-time systems Satellite images Floods Reliability

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：