检索结果-内蒙古大学图书馆

9th International Conference on Control science and Systems Engineering, ICCSSE 2023

作者： Gu, Yameng Ju, Yue Chen, Tianshi School of Data Science Shenzhen Research Institute of Big Data The Chinese Univ. of Hong Kong Shenzhen518172 China

ISBN: (纸本)9798350339055

Hysteresis system identification is a research topic in nonlinear system identification of long history. This paper proposes a novel recurrent neural network architecture to carry out hysteresis system identification. We first modify the original Prandtl-Ishlinskii (PI) model with one threshold parameter to a PI model with two threshold parameters, and then develop recurrent neural network based on the hybrid of the modified PI model and the Preisach model. We then evaluate our proposed network on systems simulated by different models, and explores the potential of model based neural networks for hysteresis system identification. As shown by numerical simulations, the proposed recurrent neural network can achieve better performances in describing different kinds of hysteresis systems. © 2023 IEEE.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

Comparative Analysis of Different Approaches to Object Detection: A Survey 3

Comparative Analysis of Different Approaches to Object Detec...

引用

3rd International Conference on Smart Electronics and Communication, ICOSEC 2022

作者： Sajjad, Mohammed Hemavathi, D. School of Computing Srm Institute of Science and Technology-Kattankulathur Department of Data Science and Business Systems Chennai India

ISBN: (数字)9781665497640

ISBN: (纸本)9781665497640

Object detection has a pivotal role in the field of security. Often, a security breach occurs under cover of night when the visibility is reduced. Manual supervision is relatively difficult as visibility is drastically reduced and human error increases to a large extent. In recent times, automation in security has gained a lot of traction. There are various inventions made in the field of automated surveillance. There are various models to detect objects in a frame with high accuracy accurately. With the help of pre-defined datasets and exclusive object detection models following the concepts of deep learning, an automated surveillance system can be implemented with ease. This study discusses the impact of automated surveillance, both night vision and daytime surveillance, the pros and cons, and the development of automated surveillance models that lead to accurate results. The study mainly focuses on deep learning concepts used to improve object detection. © 2022 IEEE.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Rényi State Entropy Maximization for Exploration Acceleration in Reinforcement Learning

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2023年第5期4卷 1154-1164页

作者： Yuan, Mingqi Pun, Man-On Wang, Dong Chinese University of Hong Kong School of Science and Engineering Shenzhen518172 China Shenzhen Research Institute of Big Data Shenzhen518172 China

One of the most critical challenges in deep reinforcement learning is to maintain the long-term exploration capability of the agent. To tackle this problem, it has been recently proposed to provide intrinsic rewards for the agent to encourage exploration. However, most existing intrinsic reward-based methods proposed in the literature fail to provide sustainable exploration incentives, a problem known as vanishing rewards. In addition, these conventional methods incur complex models and additional memory in their learning procedures, resulting in high computational complexity and low robustness. In this work, a novel intrinsic reward module based on the Rényi entropy is proposed to provide high-quality intrinsic rewards. It is shown that the proposed method actually generalizes the existing state entropy maximization methods. In particular, a k-nearest neighbor estimator is introduced for entropy estimation while a k-value search method is designed to guarantee the estimation accuracy. Extensive simulation results demonstrate that the proposed Rényi entropy-based method can achieve higher performance as compared to existing schemes. © 2020 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Oil Spill Detection Using 2D Convolution Neural Network and Generative Adversarial Network 4

Oil Spill Detection Using 2D Convolution Neural Network and ...

引用

4th International Conference on Expert Clouds and Applications, ICOECA 2024

作者： Madhura, R. Nikkath Bushra, S. Syed Ismail, A. Haleema, P.K. SRM Institute of Science and Technology Department of Computing Technologies KattankulathurTamil Nadu Chennai India SRM Institute of Science and Technology Department of Data Science and Business Systems Kattankulathur Tamil Nadu Chennai India School of Computing and Engineering University of West London RAK Branch Campus United Arab Emirates

ISBN: (纸本)9798350385793

Finding oil spills in the ocean is one of the most crucial tasks in preserving our ecosystem. Using satellite or aerial photographs as input to a deep learning model that makes use of 2D CNN (Convolution Neural Network) and GAN (Generative Adversarial Network). Deep learning assists in automatically detecting oil spills in the ocean. A strong model that may be applied to volumetric data is 2D CNN. To protect the environment and its biodiversity, it is critical to identify and minimize oil spills in delicate maritime environments. Conventional techniques for tracking oil spills in satellite imagery sometimes depend on labor-intensive, human error-prone manual inspection. When deep learning models are taught with a lot of data, they perform well. The Gathering of datasets, however the situation proved to be challenging. As a result, employing Generative Adversarial Networks (GAN) to produce new photos by utilizing preexisting ones. As a result, by giving the model a suitable number of photos to use for training and testing, it improves accuracy. © 2024 IEEE.

关键词： Satellite imagery

来源：评论

学校读者我要写书评

暂无评论

Generative Multimodal data Augmentation for Low-Resource Multimodal Named Entity Recognition 24

Generative Multimodal Data Augmentation for Low-Resource Mul...

引用

32nd ACM International Conference on Multimedia, MM 2024

作者： Li, Ziyan Yu, Jianfei Yang, Jia Wang, Wenya Yang, Li Xia, Rui School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China College of Computing and Data Science Nanyang Technological University Singapore Wee Kim Wee School of Communication and Information Nanyang Technological University Singapore

ISBN: (纸本)9798400706868

As an important task in multimodal information extraction, Multimodal Named Entity Recognition (MNER) has recently attracted considerable attention. One key challenge of MNER lies in the lack of sufficient fine-grained annotated data, especially in low-resource scenarios. Although data augmentation is a widely used technique to tackle the above issue, it is challenging to simultaneously generate synthetic text-image pairs and their corresponding high-quality entity annotations. In this work, we propose a novel Generative Multimodal data Augmentation (GMDA) framework for MNER, which contains two stages: Multimodal Text Generation and Multimodal Image Generation. Specifically, we first transform each annotated sentence into a linearized labeled sequence, and then train a Label-aware Multimodal Large Language Model (LMLLM) to generate the labeled sequence based on a label-aware prompt and its associated image. We further employ a Stable Diffusion model to generate the synthetic images that are semantically related to these sentences. Experimental results on three benchmark datasets demonstrate the effectiveness of the proposed GMDA framework, which consistently boosts the performance of several competitive methods for two subtasks of MNER in both full-supervision and low-resource settings. The low-resource dataset and source code are released at https://***/NUSTM/GMDA. © 2024 ACM.

关键词： data augmentation

来源：评论

学校读者我要写书评

暂无评论

Clear Text Visibility: A CNN-LSTM Approach for Frame of Interest Identification in Book Flipping Videos

Clear Text Visibility: A CNN-LSTM Approach for Frame of Inte...

引用

2024 IEEE International Conference on Modeling, Simulation and Intelligent computing, MoSICom 2024

作者： Buddhawar, Gaurav Dave, Dharin Jariwala, Krupa Chattopadhyay, Chiranjoy Sardar Vallabhbhai National Institute of Technology Dept. of Computer Science and Engineering Surat India Pune India Flame University School of Computing and Data Sciences Pune India

ISBN: (纸本)9798331533311

Book flipping videos present a distinctive challenge for information extraction, requiring the identification of frames with clear text visibility during dynamic page turns. This paper introduces a novel approach to frame classification in book flipping videos, leveraging the combined effect of Convolutional Neural Networks (CNNs) and Long Short-Term Memory networks (LSTMs). In particular, a CNN-LSTM model is employed, where a pre-trained CNN captures spatial features, and LSTMs extend the model's capability to distinguish temporal dependencies critical for page turn detection. Unlike conventional methods, this approach processes entire video sequences, enabling the model to learn intricate Frame of Interest (FoI) detection features indicative of page turns and text visibility. The proposed model offers improved interpretability, effectively classifying frames without compromising transparency. The research opens avenues for automated book digitization, digital library creation, and educational technology applications, enhancing information extraction and reliability from book flipping videos. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

AccelES: Accelerating Top-K SpMV for Embedding Similarity via Low-bit Pruning 31

AccelES: Accelerating Top-K SpMV for Embedding Similarity vi...

引用

31st IEEE International Symposium on High Performance Computer Architecture, HPCA 2025

作者： Zhai, Jiaqi Shi, Xuanhua Huang, Kaiyi Ye, Chencheng Hu, Weifang He, Bingsheng Jin, Hai Huazhong University of Science and Technology National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Wuhan430074 China National University of Singapore School of Computing 119077 Singapore

ISBN: (纸本)9798331506476

In the realm of recommendation systems, achieving real-time performance in embedding similarity tasks is often hindered by the limitations of traditional Top-K sparse matrix-vector multiplication (SpMV) methods, which suffer from high latency due to inefficient memory access patterns. This paper identifies these critical gaps and introduces AccelES, a novel approach that significantly enhances the efficiency of Top-K SpMV. Our method employs a two-stage calculation scheme: the first stage utilizes a compact, low-bit dataset to quickly identify the most relevant entries, while the second stage performs full-precision calculations solely on this pruned subset, thereby minimizing computational overhead. Furthermore, AccelES incorporates innovative matrix representations, Ultra-CSR and Random-CSR, which optimize memory bandwidth utilization. Experimental results demonstrate that AccelES accelerates performance, surpassing state-of-the-art FPGA, GPU, and CPU solutions by factors of 3.4×, 2.5×, and 153.3×, respectively, under controlled conditions. These advancements not only enhance processing speed but also significantly improve real-time performance in recommendation systems, establishing AccelES as a pivotal contribution to the field of Top-K sparse matrix-vector multiplication. © 2025 IEEE.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

A TRUST-REGION METHOD FOR NONSMOOTH NONCONVEX OPTIMIZATION

引用

Journal of Computational Mathematics 2023年第4期41卷 683-716页

作者： Ziang Chen Andre Milzarek Zaiwen Wen Department of Mathematics Duke UniversityUSA School of Data Science(SDS) The Chinese University of Hong KongShenzhenShenzhen Research Institute for Big Data(SRIBD)China Beijing International Center for Mathematical Research Peking UniversityBeijing 100000China

We propose a trust-region type method for a class of nonsmooth nonconvex optimization problems where the objective function is a summation of a(probably nonconvex)smooth function and a(probably nonsmooth)convex *** model function of our trust-region subproblem is always quadratic and the linear term of the model is generated using abstract descent ***,the trust-region subproblems can be easily constructed as well as efficiently solved by cheap and standard *** the accuracy of the model function at the solution of the subproblem is not sufficient,we add a safeguard on the stepsizes for improving the *** a class of functions that can be“truncated”,an additional truncation step is defined and a stepsize modification strategy is *** overall scheme converges globally and we establish fast local convergence under suitable *** particular,using a connection with a smooth Riemannian trust-region method,we prove local quadratic convergence for partly smooth functions under a strict complementary *** numerical results on a family of Ei-optimization problems are reported and demonstrate the eficiency of our approach.

关键词： Trust-region method Nonsmooth composite programs Quadratic model function Global and local convergence

来源：评论

学校读者我要写书评

暂无评论

On Kernel Design for Non-Causal Systems with Application to Feedforward Control 41

On Kernel Design for Non-Causal Systems with Application to ...

引用

第41届中国控制会议

作者： Xiaozhu Fang Tianshi Chen School of Data Science and Shenzhen Research Institute of Big Data The Chinese University of Hong Kong

ISBN: (数字)9789887581536

ISBN: (纸本)9781665482561

The stable inversion technique is widely used to design feedforward controllers of non-minimum phase ***,the stable-inversion-based feedforward controller can be identified from data in the form of a non-causal finite impulse response *** kernel-based identification of this model has been studied in previous research,and some non-causal kernels have been *** this paper,we revisit the non-causal kernels and mainly explore the design of off-diagonal *** particular,we design a new kernel by using the system theory method,which employs the multiplicatuve uncertainty *** advantage of the new kernel is confirmed in a simulation case study.

关键词： Feedforwad control Stable inversion System identification Kernel method Multiplicative uncertainty

来源：评论

学校读者我要写书评

暂无评论

Gesture Showdown: Rock Paper Scissors with AI Vision

Gesture Showdown: Rock Paper Scissors with AI Vision

引用

2025 IEEE International Conference on Computational, Communication and Information Technology, ICCCIT 2025

作者： Hemal Babu, H. Yeshwanth, C.D. Shravan, N. Bose, K. Regin Rajalakshmi Institute of Technology Department of Artificial Intelligence and Data Science Chennai India Rajalakshmi Institute of Technology Department of AI&DS and Cloud Computing Chennai India

ISBN: (纸本)9798331512965

This paper presents a gesture-based RockPaper-Scissors game that leverages computer vision and hand gesture recognition to enable intuitive human-computer interaction. Utilizing MediaPipe's pre-trained hand tracking module, the system detects and identifies 21 hand landmarks in real-time to classify gestures as rock, paper or scissors based on angular relationships between key points. OpenCV is used for webcam integration and video frame preprocessing, ensuring seamless interaction. The game compares the recognized gesture with a computer-generated move, dynamically displaying the results and resetting after each round for continuous play. The system identifies gestures through precise angle calculations, differentiating between closed fists, open palms and extended fingers. Sample outputs demonstrate real-time gameplay scenarios, highlighting gesture recognition accuracy and user engagement. Future enhancements include refining recognition accuracy, adding multiplayer functionality, introducing customizable gestures, and deploying the game on web and mobile platforms for broader accessibility. This project showcases the potential of integrating AI and computer vision in creating engaging and user-friendly applications. © 2025 IEEE.

关键词： Gesture recognition

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：