检索结果-内蒙古大学图书馆

20th CSI International Symposium on Artificial Intelligence and Signal Processing, AISP 2024

作者： Ghaffari, Shervin Yousefimehr, Behnam Ghatee, Mehdi University of Technology Dept. of Math. & Computer Sci. Amirkabir Tehran Iran

ISBN: (纸本)9798350383942

In recent years, the advancement of generative AI has profoundly influenced its application across various industries. One such industry is e-commerce where this technology can enhance both customer experience as well as merchants' productivity and profitability. In this paper, our objective is to review some of the potential use cases of generative AI technology in different areas of an online store. Specifically, we will focus on the use cases of product description generation, sentiment analysis of product reviews, and product tagging and categorization. We utilize various prompt engineering techniques to suggest exemplary implementations of these applications using large language models (LLMs). Lastly, the paper will discuss the possible risks and challenges that come with using generative AI in these contexts. © 2024 IEEE.

关键词： Sentiment analysis

来源：评论

学校读者我要写书评

暂无评论

Bilinear Hyperspectral Unmixing via Tensor Decomposition 31

Bilinear Hyperspectral Unmixing via Tensor Decomposition

引用

31st European Signal Processing Conference, EUSIPCO 2023

作者： Ding, Meng Fu, Xiao Zhao, Xi-Le School of Math. Southwest Jiaotong Univ. Chengdu China School of Elec. Eng. & Computer Sci. Oregon State Univ. CorvallisOR97331 United States School of Math. Sci. Univ. of Elec. Sci. & Tech. of China Chengdu China

ISBN: (纸本)9789464593600

The bilinear mixture model (BMM) is often used in hyperspectral unmixing (HU) for incorporating nonlinear effects, which is considered more general than the widely used linear mixture model (LMM). Existing BMM-based HU methods often lack identifiability guarantees of the endmembers and their abundances unless some stringent conditions are met. This work puts forth a new framework for BMM-based HU. Our method models the hyperspectral image as a latent factor-structured block-term tensor decomposition model with multilinear rank-(Lr, Lr, 1) terms ("LL1" model for short). This way, the HU task boils down to finding the block terms of the tensor model. The LL1 model's essential uniqueness naturally provides identifiability guarantees of the endmembers/abundances under reasonably mild conditions. An alternating gradient projection (GP) algorithm is proposed to takle the formulated tensor decomposition-based BMM-HU problem. Simulations on semi-real and real datasets show the high-quality unmixing performance of the proposed GP algorithm compared to state-of-the-art methods. © 2023 European Signal Processing Conference, EUSIPCO. All rights reserved.

关键词： Tensors

来源：评论

学校读者我要写书评

暂无评论

Generative-AI in E-Commerce: Use-Cases and Implementations

Generative-AI in E-Commerce: Use-Cases and Implementations

引用

International Symposium on Artificial Intelligence and Signal Processing (AISP)

作者： Shervin Ghaffari Behnam Yousefimehr Mehdi Ghatee Dept. of Math. & Computer Sci. Amirkabir University of Technology Tehran Iran

In recent years, the advancement of generative AI has profoundly influenced its application across various industries. One such industry is e-commerce where this technology can enhance both customer experience as well as merchants’ productivity and profitability. In this paper, our objective is to review some of the potential use cases of generative AI technology in different areas of an online store. Specifically, we will focus on the use cases of product description generation, sentiment analysis of product reviews, and product tagging and categorization. We utilize various prompt engineering techniques to suggest exemplary implementations of these applications using large language models (LLMs). Lastly, the paper will discuss the possible risks and challenges that come with using generative AI in these contexts.

关键词：

来源：评论

学校读者我要写书评

暂无评论

BCNN: A Bantamweight Convolutional Neural Network for P300 Detection

BCNN: A Bantamweight Convolutional Neural Network for P300 D...

引用

2023 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2023

作者： Liu, M. Wang, S. Beyette, F.R. University of Georgia School of Electrical and Computer Engineering AthensGA United States Hebei Normal Univ. of Sci. & Tech Sch. of Math. and Info. Sci. & Tech Hebei Qinhuangdao China Yanshan University School of Information Science and Engineering Hebei Qinhuangdao China

ISBN: (纸本)9798350341256

As one kind of Event-Related Potentials (ERPs), P300 plays an important role in studying neural activities and cognitive processes, and lays the foundation for P300 speller - the Brain computer Interface (BCI) working by detecting P300. However, due to P300's high subject variability and low Signal Noise Ratio (SNR), it's always a challenge to do P300 detection. Convolutional Neural Networks (CNNs) have been widely recognized as effective methods for P300 detection, leading to the development of increasingly complex CNN architectures for P300 detection in recent years. However, this paper takes a different approach by proposing the bantamweight CNN (BCNN), the simplest and fast-training CNN for P300 detection. BCNN comprises only one convolutional filter and a total of 141 parameters. Surprisingly, it achieves state-of-the-art performance after just 2 epochs of training, making it an exceptionally lightweight and fast CNN for P300 detection. BCNN not only provides new ideas for CNN design, but also effectively address resource and time constrained situations. © 2023 IEEE.

关键词： Brain computer interface

来源：评论

学校读者我要写书评

暂无评论

Recent Progress in Understanding Transformer and Developing Its Surpasser: A Survey

Recent Progress in Understanding Transformer and Developing ...

引用

2024 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2024

作者： Liu, M. Yang, D. Beyette, F.R. Wang, S. Zhao, L. University Of Georgia School Of Electrical And Computer Engineering AthensGA United States University Of Georgia Institute For Artificial Intelligence AthensGA United States Hebei Normal University Of Sci. & Tech. School Of Math. And Info. Tech. Qinhuangdao China

ISBN: (纸本)9798350388572

Transformers have significantly impacted machine learning, particularly in natural language processing and computer vision, due to their robust attention mechanisms and scalability. Although it is very successful and has the attention mechanism to explain it, it is still a black box in general. This lack of transparency is not conducive to applications, nor is it conducive to the development of more advanced models. We systematically analyze current methodologies for interpreting the attention patterns, hidden representations, and decision-making processes within Transformers. Additionally, we investigate how these insights aid in refining Transformer architectures and inspire the creation of innovative models that extend beyond the conventional Transformer frameworks. Due to its expertise in long-term dependencies and data scalability, Transformer undoubtedly has great advantages in foundation language models, achieving multimodality, and processing discrete data, but it is inferior to Mamba in terms of efficiency, especially when processing continuous data. It is inferior to Diffusion Model when processing image data because it is not good at grasping the distribution of high-dimensional data and is relatively less suitable for image generation model. By integrating achievements in both explaining and advancing Transformer-like architectures, this paper serves as a valuable resource for researchers aiming to enhance the performance, transparency, and efficiency of Transformers in various applications and to develop models that surpass current Transformer paradigms. © 2024 IEEE.

关键词： Scalability

来源：评论

学校读者我要写书评

暂无评论

Bilinear Hyperspectral Unmixing via Tensor Decomposition

Bilinear Hyperspectral Unmixing via Tensor Decomposition

引用

European Signal Processing Conference (EUSIPCO)

作者： Meng Ding Xiao Fu Xi-Le Zhao School of Math. Southwest Jiaotong Univ. Chengdu China School of Elec. Eng. & Computer Sci. Oregon State Univ. Corvallis OR United States School of Math. Sci. Univ. of Elec. Sci. & Tech. of China Chengdu China

The bilinear mixture model (BMM) is often used in hyperspectral unmixing (HU) for incorporating nonlinear effects, which is considered more general than the widely used linear mixture model (LMM). Existing BMM-based HU methods often lack identifiability guarantees of the endmembers and their abundances unless some stringent conditions are met. This work puts forth a new framework for BMM-based HU. Our method models the hyperspectral image as a latent factor-structured block-term tensor decomposition model with multilinear rank $(L_{r},\ L_{r},\ 1)$ terms (“LL 1” model for short). This way, the HU task boils down to finding the block terms of the tensor model. The LL 1 model's essential uniqueness naturally provides identifiability guarantees of the endmembers/abundances under reasonably mild conditions. An alternating gradient projection (GP) algorithm is proposed to takle the formulated tensor decomposition-based BMM-HU problem. Simulations on semi-real and real datasets show the high-quality unmixing performance of the proposed GP algorithm compared to state-of-the-art methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A New Clustering Approach based on Trust and Rat Swarm Algorithm for WSN Applications

A New Clustering Approach based on Trust and Rat Swarm Algor...

引用

2023 International Conference on Advances in Intelligent Computing and Applications, AICAPS 2023

作者： Osamy, Walid Salim, Ahmed Ali, Amel Al Khedr, Ahmed M. Faculty of Computers and Artificial Intelligence Benha University Computer Science Department Benha Egypt Zagazig University Math. Department Faculty of Science Zagazig Egypt Qassim University Department of Cs College of Sci. and Arts Al-Methnab Saudi Arabia Sharjah University Dept. of Information Systems Sharjah27272 United Arab Emirates Sharjah University Dept. of Computer Science Sharjah27272 United Arab Emirates

ISBN: (纸本)9798350333817

In this paper, a clustering approach called CATRSO is proposed. The selection of cluster heads (CH) is performed by considering the trust value of the nodes in order to select the most trustworthy nodes as CH and Rat Swarm Optimizer is employed for CH selection process. The trust value of the nodes and remaining energy are taken into account while designing the fitness function. In addition, a chain routing approach is employed between CHs for energy savings. The results demonstrate that the CATRSO technique is successful in selecting the most trustworthy nodes as CH and outperforms earlier efforts in the literature in terms of energy efficiency, average network lifetime, and trustworthiness of selected CHs. © 2023 IEEE.

关键词： Rats

来源：评论

学校读者我要写书评

暂无评论

On Vertex-Euclidean Deficiency of Complete Fan Graphs and Complete Wheel Graphs

引用

Journal of Combinatorial math.matics and Combinatorial Computing 2024年 121卷 41-52页

作者： Gao, Zhen-Bin Guo, Ricky Kwong, Harris Lee, Sin-Min Qiu, Wei College of General Education Guangdong University of Science and Technology Dongguan523000 China Dept. of Computer Science Univ. of Calif. at Los Angeles Los AngelesCA90095 United States Dept. of Math. Sci. SUNY Fredonia FredoniaNY14063 United States 1786 Plan Tree Drive UplandCA91784 United States College of Math. Sci.Harbin Engineering Univ. Harbin150001 China

A simple graph G with p vertices is said to be vertex-Euclidean if there exists a bijection f : V(G) → {1, 2, . . ., p} such that f(v1) + f(v2) > f(v3) for each C3-subgraph with vertex set {v1, v2, v3}, where f(v1) 2) 3). More generally, the vertex-Euclidean deficiency of a graph G is the smallest integer k such that G ∪ Nk is vertex-Euclidean. To illustrate the idea behind this new graph labeling problem, we study the vertex-Euclidean deficiency of two new families of graphs called the complete fan graphs and the complete wheel graphs. We also explore some related problems, and pose several research topics for further study. © 2024 the Author(s)

关键词： Wheels

来源：评论

学校读者我要写书评

暂无评论

Two Numerical ways RO (SuMM) and RO (MSuSu)) to Evaluate triple integrals 1

Two Numerical ways RO (SuMM) and RO (MSuSu)) to Evaluate tri...

引用

1st Iraqi Academics Syndicate International Conference for Pure and Applied sci.nces, IICPS 2020

作者： Aljassas, Safaa M. Hanoon, Wafaa H. Showq Mohammed, E. Dep. of Math. College of Edu. for Girls Univ. of Kufa Iraq Dept. of Computer Sci. College of Edu. for Girls Uuiv. of Kufa Iraq

The main goal of this research to calculate the triple integrals which has continuous functions numerically by two methods are RO(SuMM) and RO(SuMM)that obtained by Romberg acceleration with two composite rules. The first rule is the Mid-point method on the first dimension X and the second dimension Y with a suggested method on the third dimension Z and, which is denoted by SuMM. The second rule is a suggested method on the first dimension X and the second dimension Y with the Mid-point method on the third dimension Z and, which is denoted by MSuSu. Where the number of divisions on these three dimensions is exactly the Sam. We represented two theorems with their proofs to find such rules and the correction terms (the error terms) for each of rule Therefore, we obtained high accuracy results in a relatively few sub intervals and short time. Subject Classification: 65XX © Published under licence by IOP Publishing Ltd.

关键词： Numerical methods

来源：评论

学校读者我要写书评

暂无评论

Recent Progress in Understanding Transformer and Developing Its Surpasser: A Survey

Recent Progress in Understanding Transformer and Developing ...

引用

IEEE Signal Processing in Medicine and Biology Symposium (SPMB)

作者： M. Liu D. Yang F. R. Beyette S. Wang L. Zhao School of Electrical and Computer Engineering University of Georgia Athens Georgia USA Institute for Artificial Intelligence University of Georgia Athens Georgia USA School of Math. and Info. Tech. Hebei Normal University of Sci. & Tech. Qinhuangdao China

ISBN: (数字)9798350388572

ISBN: (纸本)9798350388589

关键词： Surveys Attention mechanisms Current transformers Image synthesis Scalability Biological system modeling Refining computer architecture Signal processing Natural language processing

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：