检索结果-内蒙古大学图书馆

5th IEEE Global Conference for Advancement in Technology, GCAT 2024

作者： Nagaveni, L.G. Nandyal, Shravya B. Shetty, Skanda S. Raj, Smriti Dayalan, Geetha Pes University Department of Computer Science and Engineering Bangalore India

ISBN: (纸本)9798350376685

Swar is a decentralized music platform, using blockchain and machine learning to change music industry fundamentals, where artists and users will reap the most benefits. Unlike other platforms, Swar offers transparent royalty payment system which is based on streaming data and has dynamic adjustments, thus, contributors are assured of fair remunerations. Through its DApp solution, Swar provides the artists with an opportunity to upload songs, get the allocated royalties using smart contracts, and store the content in the IPFS securely. ML-Model of the platform forecasts streaming numbers and notify artists about NFT creation with regard to the expected popularity and distributes royalties that helps artists make decisions on their work with confidence. On the top of that Swar comes up with Timebound NFTs, that empowers artists to provide special content and experiences to the fans which can be sold on the dedicated marketplace. Intuitive personalized song recommendation system also offers user tailored suggestions including the songs they like the most. Swar aspires to have a music ecosystem which is seen as transparent, fair and engaging to both artists and listeners. © 2024 IEEE.

关键词： Decentralized finance

来源：评论

学校读者我要写书评

暂无评论

Deciphering Deception-LSTM Combats Deepfake Audios 4

Deciphering Deception-LSTM Combats Deepfake Audios

引用

4th International Conference on Technological Advancements in Computational sciences, ICTACS 2024

作者： Acharya, Yushsvi Tyagi, Jahanvi Jangid, Mahesh Manipal University Department of Computer Science and Engineering Jaipur India

ISBN: (纸本)9798350387490

It's safe to say that the proliferation of synthetically generated media, particularly Deepfakes, has raised significant concerns regarding privacy, security, and integrity of digital content. There is a lot of research to detect these deep-fakes, however we believe audio deepfake detection is an ever evolving technology which needs to conquer deepfake generation techniques. In this paper, we propose a Long Term Short Memory model to classify real and fake audios and discuss how different hyperparameters affect the results. Four models have been tested with four different optimisers each to conclude the best model and optimiser for deepfake audio detection. © 2024 IEEE.

关键词： Fake detection

来源：评论

学校读者我要写书评

暂无评论

Semantic Segmentation with Attention Dense U-Net for Lung Extraction from X-ray Images 6

Semantic Segmentation with Attention Dense U-Net for Lung Ex...

引用

6th International Conference on Electrical engineering and Information and Communication Technology, ICEEICT 2024

作者： Mahmud Auvy, Akib Al Zannah, Rafiatul Mahbub-E-Elahi Sharif, Shezhan Mahmud, Washik Al Noor, Jannatun Research Group Department of Computer Science and Engineering Dhaka Bangladesh

ISBN: (纸本)9798350385779

Deep learning and digital image processing are crucial in medical imaging research. Lung segmentation is particularly challenging, demanding accurate differentiation of complex structures, sophisticated algorithms, and deep learning models for reliable results. In our research, We used the Shen-zhen chest X-ray dataset, comprising 566 frontal chest X-rays focused on pulmonary tuberculosis. This dataset, compiled by the Guangdong Medical College and Shenzhen No. 3 People's Hospital, was released by the United States National Library of Medicine. Images, captured using a Philips DR Digital Diagnostic system, were resized from $3000 \times 3000$ pixels to $512 \times 512$ pixels for computer-aided diagnosis research. We initially used the Attention Dense U-Net to extract lungs from chest X-rays, leveraging attention mechanisms to focus on relevant features and dense blocks to improve feature reuse. Our goal is to assess its performance for chest X-ray segmentation, comparing it with other deep learning models due to limited research in this area. So, the model is compared with four other variants of U-Net architectures to segment lung pixels from an X-ray image. After implementing the approach using a benchmark dataset and comparing it to other existing architectures, we present that Attention Dense U-Net gives the best accuracy for all given parameters, with a result of accuracy: 97.48%, Dice coefficient: 94.87%, and IoU: 93.87%. © 2024 IEEE.

关键词： Medical imaging

来源：评论

学校读者我要写书评

暂无评论

Improved Fake News Detection by Combining Sentence Transformers, Variational Autoencoders, and Topic Modelling: the VAE-Topic Model Fusion Method 5

Improved Fake News Detection by Combining Sentence Transform...

引用

5th IEEE International Conference on Computing, Power, and Communication Technologies, IC2PCT 2024

作者： Tufchi, Shivani Yadav, Ashima Ahmed, Tanveer Bennett University Computer Science and Engineering Greater Noida India

ISBN: (纸本)9798350383522

In this study, we present a pioneering method aimed at discerning the veracity of news content by leveraging state- of-the-art embedding techniques. Recognizing the importance of capturing semantic nuances, we employ sentence transformers to generate embeddings that encapsulate the essence of the news content. This process results in high-dimensional data, which, although rich in information, can be computationally intensive for subsequent analysis. To address this, we introduce a Variational Autoencoder (VAE) to achieve an effective dimensionality reduction while preserving critical information. However, given the intricate and hierarchical nature of news content, we integrate a topic modeling approach known as Pachinko Allocation Model (PAM) into our framework. PAM offers a more refined data representation by capturing the latent hierarchical relationships between various news topics, adding an extra layer of semantic understanding to our method. The enhanced data representations from the VAE and PAM processes are con- catenated, creating a comprehensive feature set. This combined feature set is then used to train multiple classifiers to segregate real news from fake counterparts. Through this comprehensive numerical investigation, we have achieved an accuracy of 96.80% and an F1 score of 97% using the transformer-based sentence paraphrase- multilingual architecture. Our study's findings are not only significant in terms of accuracy but also underscore the potential of combining multiple advanced techniques to tackle the fake news challenge. © 2024 IEEE.

关键词： Autoencoders Fake news Machine Learning Models Topic Modelling Transformers Embeddings

来源：评论

学校读者我要写书评

暂无评论

Favourable subpopulation migration strategy for travelling salesman problem

引用

International Journal of Business Intelligence and Data Mining 2022年第3期20卷 299-326页

作者： Chandar, Abhishek Srinivasan, Akshay Anand, G. Paavai Department of Computer Science and Engineering SRM Institute of Computer Science and Technology Tamil Nadu Chennai600026 India

Most of the existing migration strategies are complex in nature which results in high computation time. Since the introduction of parallel genetic algorithms, they are known for obtaining better optimal solutions at the cost of more computation. Migration has played a significant role in maintaining genetic diversity of the population. This paper aims to propose a new migration strategy that improves the performance of the parallel genetic algorithm in solving NP hard problems. In order to assess its behaviour, various travelling salesman problem datasets are considered since it is a perfect example of an NP hard problem. The algorithm is based on an island model approach where one subpopulation is nurtured by the other subpopulations by taking turns during the migration process to progressively reach the optimal solution. In turn, the nurtured subpopulation provides a feedback loop to the subpopulation with poor fitness value based on certain criteria to be met. An extensive study was conducted to understand the optimal parameters values in which the proposed algorithm works best. © 2022 Inderscience Enterprises Ltd.

关键词： Traveling salesman problem

来源：评论

学校读者我要写书评

暂无评论

ARR: An Attention-Based Robust Residual Network Architecture 10

ARR: An Attention-Based Robust Residual Network Architecture

引用

10th IEEE Smart World Congress, SWC 2024

作者： Li, Yunlong Xu, Qingguo Chen, Shengbo Zheng, Fei Qiang, Guoquan Shanghai University School of Computer Engineering and Science Shanghai China

ISBN: (纸本)9798331520861

As controllers of safety-critical systems, the robustness of feedforward neural networks is crucial. Although current research mainly focuses on developing adversarial training methods, the impact of network structure on robustness cannot be ignored. In this paper, we propose a novel robust deep network, AttentiveRobustResNet(ARR), which integrates the residual network architecture and the multi-head self-attention mechanism to effectively improve the model's robustness in the face of adversarial samples without increasing additional model parameters. By using the multi-head self-attention mechanism to replace part of the convolution operation for feature extraction, the global information capture ability of the network is improved. By adjusting the sub-components of the network (activation function, data augmentation), the impact of abnormal phenomena such as gradient disappearance and global information loss in adversarial training on the robustness of adversarial training is alleviated. Based on adopting various adversarial training methods, this paper conducts extensive experimental validation of different adversarial attacks on multiple datasets. The results show that AttentiveRobustResNet significantly improves adversarial robustness compared with other neural network models under the same parameter scale. In particular, compared to the baseline network ResNet50, our model achieves an 8 % improvement in clean accuracy and a 5% improvement in adversarial robustness (AA) under the TRADES adversarial training framework. © 2024 IEEE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Building Empathic Chabot using LSTM and NLP 4

Building Empathic Chabot using LSTM and NLP

引用

4th International Conference on Innovative Practices in Technology and Management, ICIPTM 2024

作者： Jeyalakshmi, Jeyalakshmi Srinidhi, S. Balaprathap, C. Department of Computer Science and Engineering Amrita VishwaVidhyapeetham Chennai India

ISBN: (数字)9798350307757

ISBN: (纸本)9798350307757

This research work presents a Empathic chatbot leveraging advanced natural language processing techniques. The chatbot employs a deep learning architecture, specifically a variant of the Transformer model, to generate human-like responses in real- time. The training dataset includes diverse conversational patterns to enhance the bot's ability to handle a wide range of user inputs. The model is fine-tuned through an iterative process, optimizing for coherence, relevance, and contextual understanding. Evaluation metrics demonstrate the effectiveness of the generative chatbot in providing contextually appropriate and coherent responses across various domains. The research contributes to the evolving field of conversational AI, showcasing the potential of generative models for creating engaging and context-aware chatbot interactions. © 2024 IEEE.

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

Dynamic L-Budget Clustering of Curves 19

Dynamic L-Budget Clustering of Curves

引用

19th Scandinavian Symposium on Algorithm Theory, SWAT 2024

作者： Buchin, Kevin Buchin, Maike Gudmundsson, Joachim Plätz, Lukas Thiel, Lea Wong, Sampson Faculty of Computer Science TU Dortmund University Germany Faculty of Computer Science Ruhr University Bochum Germany Faculty of Engineering The University of Sydney Australia Department of Computer Science University of Copenhagen Denmark

ISBN: (纸本)9783959773188

A key goal of clustering is data reduction. In center-based clustering of complex objects therefore not only the number of clusters but also the complexity of the centers plays a crucial role. We propose LBudget Clustering as unifying perspective on this task, optimizing the clustering under the constraint that the summed complexity of all centers is at most L. We present algorithms for clustering planar curves under the Fréchet distance, but note that our algorithms more generally apply to objects in metric spaces if a notion of simplification of objects is applicable. A scenario in which data reduction is of particular importance is when the space is limited. Our main result is an efficient (8 + Ε)-approximation algorithm with a (1 + Ε)-resource augmentation that maintains an L-budget clustering under insertion of curves using only O(LΕ−1) space and O∗(L3 log L + L2 log(r∗/r0)) time where O∗ hides factors of Ε−1 © Kevin Buchin, Maike Buchin, Joachim Gudmundsson, Lukas Plätz, Lea Thiel, and Sampson Wong;licensed under Creative Commons License CC-BY 4.0.

关键词： Approximation algorithms

来源：评论

学校读者我要写书评

暂无评论

A New Multi-objective Optimization Algorithm Based on Decomposition with Dynamic Resource Allocation 8

A New Multi-objective Optimization Algorithm Based on Decomp...

引用

8th International Conference on Electrical, Mechanical and computer engineering, ICEMCE 2024

作者： Zhu, Huazheng Zhang, Jun Jiang, Aiping Liu, Yaping Nie, Zicheng Chongqing University of Science & Technology School of Computer Science and Engineering Chongqing China

ISBN: (纸本)9798331506230

The Dynamic Resource Allocation Multi-Objective Optimization Algorithm (MOEA/D-DRA) is a method for solving multi-objective optimization problems (MOPs). This algorithm enhances the performance of the original MOEA/D algorithm through a strategy of dynamically allocating computational resources. However, for some challenging problems, the utility function of MOEA/D-DRA may not effectively allocate resources. This paper proposes an improved MOEA/D-DRA algorithm, introducing a new utility function to more accurately reflect the difficulty of subproblems in the current population, named MOEA/D-DRA-NEF. Experimental comparisons with ten other state-of-the-art MOEA algorithms demonstrate that MOEA/D-DRA-NEF outperforms the other algorithms in terms of convergence and diversity. © 2024 IEEE.

关键词： Optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Hybrid Decomposition Network for Light Field Super-Resolution 5

Hybrid Decomposition Network for Light Field Super-Resolutio...

引用

5th International Conference on computer Vision, Image and Deep Learning, CVIDL 2024

作者： Zheng, Wei Chen, Xiaoming Wang, Chen Wang, Xiaochuan Hu, Zeke Zexi Chung, Yuk Ying Beijing Technology and Business University School of Computer Science and Engineering Beijing China The University of Sydney School of Computer Science Sydney Australia

ISBN: (纸本)9798350373820

abstract-Light field cameras are considered emerging 3D world sensing instruments capturing dense images. High-resolution images would provide richer semantic cues than low-resolution ones to facilitate various computer vision-based tasks. The use of spatial-angular information plays a key role in light field image super-resolution (LFSR), which is an effective way to increase image resolution. A considerable amount of recent works have employed Transformer-based models in LFSR tasks and achieved performance improvement. However, most of them enforced the self-attention mechanism equally on both the light field's spatial and angular domains, without taking into account the distinct characteristics of each domain. On the one hand, the angular inter-view correlation tends to be low-frequency, where self-attention works more effectively as a low-pass filter. On the other hand, the spatial intra-view information is of high-frequency, which is more suitable for being processed by convolution layers as high-pass filters. With this insight, we propose a Hybrid Decomposition Network for LFSR (LF-HDNet) that combines convolution and self-attention but performs them separately on the two domains to extract spatial-angular features for LFSR in this paper. Experimental results demonstrate the superior performance of our method over the state-of-the-art methods in LFSR tasks. © 2024 IEEE.

关键词： Convolution

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：