检索结果-内蒙古大学图书馆

Deep Fake Detection using Transfer Learning: A Comparative study of Multiple Neural Networks

Deep Fake Detection using Transfer Learning: A Comparative s...

2024 International Conference on Signal Processing, Computation, Electronics, Power and Telecommunication, IConSCEPT 2024

作者： Abhineswari, M. Charan, Kurupati Sai Bn, Shrikarti Kanmani R, Sujithra Shool of Computer Science and Engineering Vellore Institute of Technology Chennai India

ISBN: (纸本)9798331540685

With the proliferation of sophisticated AI techniques, the creation and dissemination of deep fake images and videos have emerged as a pressing concern in today's digital landscape. Deep fake technology employs advanced machine learning algorithms to manipulate or synthesize realistic-looking media, often with malicious intent. This study evaluates the effectiveness of various pre-trained deep learning models using transfer learning for detecting deep fake images on the Face Forensics++ dataset. The models considered include MobileN etV2, ResN et50, Inception V3, EfficientN et, Xception, NASNetMobile, and a Custom CNN. Accuracies obtained from these models are compared to assess their performance in distinguishing between real and fake images. With the highest performance in MobileNetV2 with 89% followed by ResNet50 with 83 % and other models. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Efficient-Learning Grasping and Pushing in Dense Stacking via Mask Function and Pixel Overlap Rate 7

Efficient-Learning Grasping and Pushing in Dense Stacking vi...

引用

7th International Symposium on Autonomous Systems, ISAS 2024

作者： Lian, Jie Jiang, Juncheng Qiu, Chaochao Pan, Qinghui Dong, Yongxiang Wang, Zhao Wang, Dong Dalian University of Technology School of Control Science and Engineering Dalian China University of Science and Technology Liaoning Shool of Computer Science and Software Engineering China

ISBN: (纸本)9798350363173

In the field of robotic grasping, grasping tightly stacked objects is a formidable challenge. The presence of non-target objects significantly increases the risk of grasping failure. To solve this problem, the aim is to take action while grasping, reduce clutter, and effectively complete grasping tasks. Based on deep reinforcement learning, the Efficient Learning Grasping(ELG) framework is proposed that integrates pushing actions alongside grasping to efficiently pick up objects from cluttered environments with reduced training time. However, in the process of grasping, the agent is rewarded only when the grasp is successful, so the data collection is extremely inefficient and spends a lot of time on the training of grasping strategy. In our work, two strategies are proposed to improve the training efficiency. First, two mask functions are introduced as prior knowledge to enable the agent to focus on more meaningful regions while ignoring some unnecessary regions, aiming at effective learning. Second, The Pixel Overlap Rate (POR) is introduced to quantify the environment clutter level, and set a new push action reward function based on the POR. The implementation of the above ideas in a simulation environment and transferred the learned models to the real world and verified them in practice. Compared with existing method, our method improves the training efficiency of the initial algorithm while having a higher success rate of grasping. Detailed grasping video at https://***/CqtSvaVl60M. © 2024 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

VisualRWKV: Exploring Recurrent Neural Networks for Visual Language Models 31

VisualRWKV: Exploring Recurrent Neural Networks for Visual L...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Hou, Haowen Zeng, Peigen Ma, Fei Yu, Fei Richard Shenzhen China College of Computer Science and Software Engineering Shenzhen University Shenzhen China Shool of Information Technology Carleton University Canada

ISBN: (纸本)9798891761964

Visual Language Models (VLMs) have rapidly progressed with the recent success of large language models. However, there have been few attempts to incorporate efficient linear Recurrent Neural Networks (RNNs) architectures into VLMs. In this study, we introduce VisualRWKV, the first application of a linear RNN model to multimodal learning tasks, leveraging the pre-trained RWKV language model. We propose a data-dependent recurrence and sandwich prompts to enhance our modeling capabilities, along with a 2D image scanning mechanism to enrich the processing of visual sequences. Extensive experiments demonstrate that VisualRWKV achieves competitive performance compared to Transformer-based models like LLaVA-1.5 on various benchmarks. Compared to LLaVA-1.5, VisualRWKV has a speed advantage of 3.98 times and can save 54% of GPU memory when reaching an inference length of 24K tokens. To facilitate further research and analysis, we have made the checkpoints and the associated code publicly accessible at the following GitHub repository: https://***/howard-hou/VisualRWKV. © 2025 Association for Computational Linguistics.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

Deep Fake Detection using Transfer Learning: A Comparative study of Multiple Neural Networks

Deep Fake Detection using Transfer Learning: A Comparative s...

引用

Signal Processing, Computation, Electronics, Power and Telecommunication (IConSCEPT), International Conference on

作者： Abhineswari M Kurupati Sai Charan Shrikarti BN Sujithra Kanmani R Shool of Computer Science and Engineering Vellore Institute of Technology Chennai India

ISBN: (数字)9798331540685

ISBN: (纸本)9798331540692

关键词： Deepfakes Accuracy Computational modeling Transfer learning Neural networks Pressing Signal processing

来源：评论

学校读者我要写书评

暂无评论

Efficient-Learning Grasping and Pushing in Dense Stacking via Mask Function and Pixel Overlap Rate

Efficient-Learning Grasping and Pushing in Dense Stacking vi...

引用

International Symposium on Autonomous Systems (ISAS)

作者： Jie Lian Juncheng Jiang Chaochao Qiu Qinghui Pan Yongxiang Dong Zhao Wang Dong Wang School of Control Science and Engineering Dalian University of Technology Dalian China Shool of Computer Science and Software Engineering University of Science and Technology Liaoning

ISBN: (数字)9798350363173

ISBN: (纸本)9798350363180

关键词： Training Stacking Grasping Production Data collection Deep reinforcement learning Classification algorithms

来源：评论

学校读者我要写书评

暂无评论

The Design and Practice of the OpenKylin Build and Management System 3

The Design and Practice of the OpenKylin Build and Managemen...

引用

3rd IEEE International Conference on Software Engineering and Artificial Intelligence, SEAI 2023

作者： Peng, Long Liu, Xiaodong Yu, Jie Liao, Xia Tashi, Nyima Xu, Xinyan Wang, Wenzhu Rengzeng, Duojie Zhou, Mao National University of Defense Technology College of Computer Changsha China Tibet University Shool of Information Science and Technology Lhasa China Haihe Lab of Itai Tianjin China KylinSoft Co. Ltd. Department of Desktop Tianjin China

ISBN: (纸本)9798350337952

The Linux-based desktop operating systems now dominate China's Market of Information technology Application Innovation, as well as the kernel drivers and application companions. All of them can be arranged as software packages, such as Linux kernels, system configurations, core dynamic libraries. Building such a sophisticated and dynamic ecosystem urges for a clear and pipelined architecture of the build system. In this paper we elaborate our design and practice of the openKylin build and management system, including a property-extended modular package management system, an incremental upgrading and updating scheme, a remote maintenance system, and an enhanced modular code management system. The deployment results show that the openKylin build system has coherent procedure. © 2023 IEEE.

关键词： Linux

来源：评论

学校读者我要写书评

暂无评论

VisualRWKV: Exploring Recurrent Neural Networks for Visual Language Models

arXiv

引用

arXiv 2024年

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

Fruits Classification and Analysis in Hyperspectral Imaging Using Deep Convolutional Neural Networks 6th

Fruits Classification and Analysis in Hyperspectral Imaging ...

引用

6th International Conference on Advanced Informatics for Computing Research, ICAICR 2023

作者： Verma, Divyanshu Vijayarajan, V. Rajkumar, S. Ramesh Pathy, M. Shool of Computer Science and Engineering Vellore Institute of Technology Tamil Nadu Vellore632014 India School of Bio Sciences and Technology Vellore Institute of Technology Tamil Nadu Vellore632014 India

ISBN: (纸本)9783031840616

Hyperspectral imaging (HSI) emerges as a highly efficient means of information retention, particularly when leveraging the power of deep learning. The intricate task of non-destructively identifying rotten fruits is addressed through HSI, offering a non-invasive method for data collection that significantly streamlines the sorting process using advanced algorithms. Deep learning, employing neural network methodologies, facilitates the exploration of the intricate connection between spatial and spectral contexts within the data. The HSI system captures the wavelength information of refracted, reflected, and absorbed light across multiple bands. To address noise introduced by the machinery, a Standard Normal Variate (SNV) filter is applied, resulting in a refined dataset represented as a hyperspectral cube (3D). The workflow involves the utilization of an image pre-training convolutional neural network (CNN) on the collected dataset, subsequently fine-tuned to serve as a robust classifier. In the proposed model, a substantial majority of the dataset (70%) is dedicated to training the neural network, with the remaining portion (30%) reserved for the validation process. The findings demonstrate that this analysis approach, leveraging the CNN algorithm, yields remarkably accurate results (>90%) and proves to be highly effective in discerning fruit quality. The presented methodology showcases a promising synergy between hyperspectral imaging and deep learning, offering a reliable solution for non-destructive fruit quality assessment. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

From Handcrafted Features to LLMs: A Brief Survey for Machine Translation Quality Estimation

From Handcrafted Features to LLMs: A Brief Survey for Machin...

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Haofei Zhao Yilun Liu Shimin Tao Weibin Meng Yimeng Chen Xiang Geng Chang Su Min Zhang Hao Yang Shool of Computer Science and Engineering Northeastern University Shenyang China Huawei China National Key Laboratory for Novel Software Technology Nanjing University Nanjing China

ISBN: (数字)9798350359312

ISBN: (纸本)9798350359329

Machine Translation Quality Estimation (MTQE) is the task of estimating the quality of machine-translated text in real time without the need for reference translations, which is of great importance for the development of MT. After two decades of evolution, QE has yielded a wealth of results. This article provides a comprehensive overview of QE datasets, annotation methods, shared tasks, methodologies, challenges, and future research directions. It begins with an introduction to the background and significance of QE, followed by an explanation of the concepts and evaluation metrics for word-level QE, sentence-level QE, document-level QE, and explainable QE. The paper categorizes the methods developed throughout the history of QE into those based on handcrafted features, deep learning, and Large Language Models (LLMs), with a further division of deep learning-based methods into classic deep learning and those incorporating pre-trained language models (LMs). Additionally, the article details the advantages and limitations of each method and offers a straightforward comparison of different approaches. Finally, the paper discusses the current challenges in QE research and provides an outlook on future research directions.

关键词： Deep learning Learning systems Surveys Measurement Annotations Neural networks Estimation

来源：评论

学校读者我要写书评

暂无评论

Decoupling of Direction and Length for Cumulative Step Size Adaptation 17

Decoupling of Direction and Length for Cumulative Step Size ...

引用

17th International Conference on Computational Intelligence and Security, CIS 2021

作者： Zhang, Shuo Li, Zhenhua Yang, Detian Wang, Shuo Cai, Xinye School of Computer Science and Technology Nanjing University of Aeronautics and Astronautics Nanjing China Shool of Information Technology and Engineering Tianjin University of Technology and Education Tianjin China

ISBN: (纸本)9781665494892

Covariance matrix adaptation evolution strategy (CMA-ES) is a successful evolutionary algorithm for continuous optimization. The cumulative step size adaptation (CSA) is the default method for adapting the step size in CMA-ES and achieves remarkable performance in practice. It follows the idea of exploiting the correlations between search directions in consecutive generations. In this paper, we investigate the effects of the update directions and the length in CSA. We find that sampling from uniform sphere distribution can yield comparative performance to the default setting, and normalizing the update direction in CSA does not hurt the performance. Further, the CSA with actual search direction cannot perform well, and the correlation between the default direction of CSA and the actual search direction is rather small. It indicates that CSA does not accumulate the actual search directions. We argue that the step size and the covariance matrix should be updated by decoupled search information, and accumulating the actual search direction results in inferior performance due to the same update direction in two evolution paths for the step size and covariance matrix updates. © 2021 IEEE.

关键词： Covariance matrix

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：