检索结果-内蒙古大学图书馆

TRANS-VQA: Fully Transformer-Based Image Question-Answering Model Using Question-guided Vision Attention

Inteligencia Artificial 2024年第73期27卷 111-128页

作者： Koshti, Dipali Gupta, Ashutosh Kalla, Mukesh Sharma, Arvind Department of Computer Science and Engineering Sir Padampat Singhania University Rajasthan Udaipur India Department of Mathematics Sir Padampat Singhania University Rajasthan Udaipur India

Understanding multiple modalities and relating them is an easy task for humans. But for machines, this is a stimulating task. One such multi-modal reasoning task is Visual question answering which demands the machine to produce an answer for the natural language query asked based on the given image. Although plenty of work is done in this field, there is still a challenge of improving the answer prediction ability of the model and breaching human accuracy. A novel model for answering image-based questions based on a transformer has been proposed. The proposed model is a fully Transformer-based architecture that utilizes the power of a transformer for extracting language features as well as for performing joint understanding of question and image features. The proposed VQA model utilizes F-RCNN for image feature extraction. The retrieved language features and object-level image features are fed to a decoder inspired by the Bi-Directional Encoder Representation Transformer-BERT architecture that learns jointly the image characteristics directed by the question characteristics and rich representations of the image features are obtained. Extensive experimentation has been carried out to observe the effect of various hyperparameters on the performance of the model. The experimental results demonstrate that the model’s ability to predict the answer increases with the increase in the number of layers in the transformer’s encoder and decoder. The proposed model improves upon the previous models and is highly scalable due to the introduction of the BERT. Our best model reports 72.31% accuracy on the test-standard split of the VQAv2 dataset. © 2024, Asociacion Espanola de Inteligencia Artificial. All rights reserved.

关键词： Query processing

来源：评论

学校读者我要写书评

暂无评论

Unveiling the Power of Sparse Neural Networks for Feature Selection 27

Unveiling the Power of Sparse Neural Networks for Feature Se...

引用

27th European Conference on Artificial Intelligence, ECAI 2024

作者： Atashgahi, Zahra Liu, Tennison Pechenizkiy, Mykola Veldhuis, Raymond Mocanu, Decebal Constantin van der Schaar, Mihaela Faculty of Electrical Engineering Mathematics and Computer Science University of Twente Netherlands Department of Applied Mathematics and Theoretical Physics University of Cambridge United Kingdom Department of Mathematics and Computer Science Eindhoven University of Technology Netherlands Department of Computer Science University of Luxembourg Luxembourg

ISBN: (纸本)9781643685489

Sparse Neural Networks (SNNs) have emerged as powerful tools for efficient feature selection. Leveraging the dynamic sparse training (DST) algorithms within SNNs has demonstrated promising feature selection capabilities while drastically reducing computational overheads. Despite these advancements, several critical aspects remain insufficiently explored for feature selection. Questions persist regarding the choice of the DST algorithm for network training, the choice of metric for ranking features/neurons, and the comparative performance of these methods across diverse datasets when compared to dense networks. This paper addresses these gaps by presenting a comprehensive systematic analysis of feature selection with sparse neural networks. Moreover, we introduce a novel metric considering sparse neural network characteristics, which is designed to quantify feature importance within the context of SNNs. Our findings show that feature selection with SNNs trained with DST algorithms can achieve, on average, more than 50% memory and 55% FLOPs reduction compared to the dense networks, while outperforming them in terms of the quality of the selected features. © 2024 The Authors.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Neuro-Symbolic Approach for Tantrum Monitoring and Prevention in Individuals with Autism Spectrum Disorder: A Protocol for Virtual Agents 3

Neuro-Symbolic Approach for Tantrum Monitoring and Preventio...

引用

3rd International Conference on Hybrid Human-Artificial Intelligence, HHAI 2024

作者： Monaldini, Andrea Department of Computer Science University of Pisa Pisa Italy Department of Information Engineering Computer Science and Mathematics University of l'Aquila Italy

ISBN: (纸本)9781643685229

Autism Spectrum Disorder (ASD) poses unique challenges, with individuals often experiencing difficulties in emotional regulation, leading to disruptive behaviors such as tantrums. This research protocol outlines the development of an innovative system designed to monitor and prevent tantrums in individuals with ASD, employing a neuro-symbolic approach. The proposed system integrates neural networks and symbolic reasoning to improve understanding and prediction of tantrum episodes. Leveraging real-time physiological and behavioral data, collected through wearable devices and user input, the system employs machine learning algorithms to detect patterns indicative of imminent tantrum events. Furthermore, a symbolic reasoning system interprets these patterns, taking into account individualized factors. The outcomes of this study are anticipated to contribute valuable insights to the growing field of technology-assisted interventions for neurodevelopmental disorders. © 2024 The Authors.

关键词： Learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Voice assistance to blind people by detection of surrounding objects, activities, and humans

Voice assistance to blind people by detection of surrounding...

引用

2024 International Conference on Computing and Intelligent Reality Technologies, ICCIRT 2024

作者： Kambham, Suhas Preetham Varman, Raahul Ramya, Tadepalli Sri Valli Eniyan, M.K. Tangavel, Senthil Kaumar Somasundaram, K. Shanmugam, Selvanayaki Kolandapalayam Department of Computer Science and Engineering Amrita School of Computing Coimbatore India Department of Mathematics and Computer Science Ashland University United States

ISBN: (纸本)9798331510299

Various systems were developed to assist those with vision impairments and enhance their standard of life. Regrettably, the majority of these systems have restricted functionality. We decided to build a model that detects all possible surrounding things and guides them through voice description. Our main objective is to detect the obstacles in the person's pathway and warn them using voice description [1]. The model can assist blind people from the surrounding obstacles which exist in their pathway. The primary goal of the model is to identify the nearby obstacles - vehicles, animals, persons, objects, etc, in the person's pathway and guide or warn them using voice description. Our proposed system integrates advanced real-time obstacle detection and voice-guided navigation using a novel combination of DepthNet module with convolutional LSTM layers, enhancing spatial context awareness compared to traditional models. This approach, along with fine-tuning on diverse datasets such as MS COCO and BDD100K, ensures robust performance in varied environments and conditions. The real-time guidance capability is further improved by the integration of optical flow algorithms, which accurately estimate motion in dynamic scenes. This comprehensive solution significantly enhances the assistive technology for those with visual impairments. © 2024 IEEE.

关键词： Obstacle detectors

来源：评论

学校读者我要写书评

暂无评论

Leakage-Resilient Anonymous Heterogeneous Multi-Receiver Hybrid Encryption in Heterogeneous Public-Key System Settings

引用

IEEE Access 2024年 12卷 28155-28168页

作者： Tseng, Yuh-Min Tsai, Tung-Tso Huang, Sen-Shan Ho, Ting-Chieh National Changhua University of Education Department of Mathematics Changhua500 Taiwan National Taiwan Ocean University Department of Computer Science and Engineering Keelung202 Taiwan

By side-channel attacks, a fraction part of secret keys used in cryptographic schemes could be leaked to adversaries. Recently, adversaries have realized practical side-channel attacks so that these existing cryptographic schemes could be broken. Indeed, researchers have invested and proposed a good approach to withstand such attacks, called as leakage-resilient cryptography. Very recently, several leakage-resilient anonymous multi-receiver encryption (LR-AMRE) schemes based on various public-key systems were also proposed. However, these LR-AMRE schemes are not suitable for a heterogeneous public-key environment under which an authorized receiver group includes heterogeneous receivers under various PKS settings and these receivers have various types of secret/public key pairs. In this article, we propose the first leakage-resilient anonymous heterogeneous multi-receiver hybrid encryption (LR-AHMR-HE) scheme for the heterogeneous public-key system settings. A new framework and associated adversary games of the LR-AHMR-HE scheme are defined. In the adversary games, adversaries are admitted to continuously intercept a fraction part of secret keys. Under the adversary games, formal security proofs are provided to show that the proposed scheme is secure against two types of adversaries (illegitimate user and malicious authority). Comparisons with several related previous schemes are demonstrated to present the merits of our scheme. © 2013 IEEE.

关键词： Side channel attack

来源：评论

学校读者我要写书评

暂无评论

Hybrid Smart Contracts for Securing IoMT Data

引用

computer Systems science & engineering 2023年第1期44卷 457-469页

作者： D.Palanikkumar Adel Fahad Alrasheedi P.Parthasarathi S.S.Askar Mohamed Abouhawwash Department of Computer Science and Engineering Dr NGP Institute of TechnologyCoimbatore641048India Department of Statistics and Operations Research College of ScienceKing Saud UniversityRiyadh11451Saudi Arabia Department of Computer Science and Engineering Bannari Amman Institute of TechnologySathy638401India Department of Mathematics Faculty of ScienceMansoura UniversityMansoura35516Egypt Department of Computational Mathematics Scienceand Engineering(CMSE)Michigan State UniversityEast LansingMI48824USA

Data management becomes essential component of patient *** of Medical Things(IoMT)performs a wireless communication between E-medical applications and human *** of consulting a doctor in the hospital,patients get health related information remotely from the *** main issues in the E-Medical application are lack of safety,security and priv-acy preservation of patient’s health care *** overcome these issues,this work proposes block chain based IoMT Processed with Hybrid consensus protocol for secured *** health data is collected from physician,smart devices *** main goal is to store this highly valuable health related data in a secure,safety,easy access and less cost-effective *** this research we combine two smart contracts such as Practical Byzantine Fault Tolerance with proof of work(PBFT-PoW).The implementation is done using cloud technology setup with smart contracts(PBFT-PoW).The accuracy rate of PBFT is 90.15%,for PoW is 92.75%and our proposed work PBFT-PoW is 99.88%.

关键词： PoW byzantine fault tolerance IoMT cloud computing health care data

来源：评论

学校读者我要写书评

暂无评论

Convolutional neural network-based image watermarking using discrete wavelet transform

引用

International Journal of Information Technology (Singapore) 2023年第4期15卷 2021-2029页

作者： Tavakoli, Alireza Honjani, Zahra Sajedi, Hedieh Department of Computer Engineering School of Electrical and Computer Engineering College of Engineering University of Tehran Tehran Iran Department of Computer Science School of Mathematics Statistics and Computer Science College of Science University of Tehran Tehran Iran

With the growing popularity of the Internet, digital images are used and transferred more frequently. Although this phenomenon facilitates easy access to information, it also creates security concerns and violates intellectual property rights by allowing illegal use, copying, and digital content theft. Using watermarks in digital images is one of the most common ways to maintain security. Watermarking is proving and declaring ownership of an image by adding a digital watermark to the original image. Watermarks can be either text or an image placed overtly or covertly in an image and are expected to be challenging to remove. This paper proposes a combination of convolutional neural networks (CNNs) and wavelet transforms to obtain a watermarking network for embedding and extracting watermarks. The network is independent of the host image resolution, can accept all kinds of watermarks, and has only 11 layers while keeping performance. Performance is measured by two terms;the similarity between the extracted watermark and the original one and the similarity between the host image and the watermarked one. © 2023, The Author(s), under exclusive licence to Bharati Vidyapeeth's Institute of computer Applications and Management.

关键词： Convolutional neural networks Neural networks Watermarking Wavelet transform

来源：评论

学校读者我要写书评

暂无评论

Consecutive k-out-of-n:W system signature reliability appraisal via structure–function approach and u-function approach

引用

Life Cycle Reliability and Safety engineering 2024年第3期13卷 231-242页

作者： Sadiya Kumar, Akshay Ram, Mangey Department of Mathematics Graphic Era Deemed to be University Depratment of Mathematics Graphic Era Hill University Department of Mathematics Computer Science and Engineering Graphic Era Deemed to be University

The signature reliability of the consecutive k-out-of-n:W system is studied using two different approaches: the structure–function approach and the u-function approach. The connection is supposed to comprise n total components, where only k components are in a functioning or working state. Firstly, a consecutive 3-out-of-5:W system is modelled to estimate the reliability function of the system via polynomial function. Secondly, the tail-signature of the consecutive 3-out-of-5 system has been evaluated, which would allow the signature reliability evaluation. Thirdly, the determination of the degree of reliability, by using Barlow-Proschan index of the system is evaluated. Lastly, the minimal signature, predicted lifespan, and actual cost rate are derived for the proposed system. The purpose of this work is to evaluate all these outcomes and then compare their values, which are yielded by both applied approaches, i.e., the universal generating function approach and the structure–function approach. © The Author(s), under exclusive licence to Society for Reliability and Safety (SRESA) 2024.

关键词： Consecutive k-out-of-n:W system Cumulative signature Predicted lifespan Signature reliability Structure–function approach u-Function approach

来源：评论

学校读者我要写书评

暂无评论

Experimental and Statistical Study on Mechanical Behavior of H13 Tool Steel Additively Manufactured and Laser Surface Treated

引用

Lasers in Manufacturing and Materials Processing 2024年第2期11卷 193-217页

作者： Baali, Lamya Dehghan, Shayan Iltaf, Asim Barka, Noureddine Dassylva-Raymond, Véronique Department of Mathematics Computer Science and Engineering University of Quebec at Rimouski RimouskiQCG5L 3A1 Canada

H13 tool steel is used in hot work tooling applications where favorable mechanical properties and surface quality are required. This material has high hardness at elevated temperatures, high wear resistance, and good toughness. The potential for complex mold designs and conformal cooling channels using additive manufacturing technology makes it an attractive option for tool steel fabrication. This study uses the SLM technology to fabricate H13 tool steel, with laser power set at 125 W, scanning speed at 800 mm/s, layer thickness at 40 mm, and hatch space at 0.12 mm. After the SLM process, a specific heat treatment is carried out, including stress relieving, austenitization, quenching, and tempering. Laser surface heat treatment is performed to control laser power, scanning speed, and defocus distance. The results showed that the ultimate tensile strength (UTS) is improved compared to heat-treated samples, but the elongation is reduced. Fractography analysis applied by SEM revealed a combination of ductile and brittle fracture behaviors in different zones. Finally, the findings of this work contributed to the improvement of laser treatment parameters and the enhancement of mechanical properties of H13 tool steel molds. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2023.

关键词： Tensile strength

来源：评论

学校读者我要写书评

暂无评论

Adding a constant to the variables in the regression through the origin: the effect on the uncentered R²

引用

Quality and Quantity 2023年第3期57卷 2781-2789页

作者： Triacca, Umberto Department of Computer Engineering Computer Science and Mathematics University of L’Aquila Coppito 67010 Italy

It is well known the effect on uncentered R2 stemming from adding a constant to dependent variable in a linear regression model with intercept. In this paper, we investigate the effect of adding a constant to variable... 详细信息

It is well known the effect on uncentered R² stemming from adding a constant to dependent variable in a linear regression model with intercept. In this paper, we investigate the effect of adding a constant to variables on the uncentered R² when a linear regression through the origin is used. In particular, we consider two cases. First, a constant c∈ R is added to all observations of the dependent variable. Second, a constant c∈ R is added to all the observations of both the dependent variable and at least an independent variable. We show that in both cases there is an artificial variation of the uncentered R². This quantity is not invariant under location change. © 2022, The Author(s).

关键词： Coefficient of determination Least squares Linear regression model Measure of goodness of fit

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：