检索结果-内蒙古大学图书馆

8th International Conference on Communication and Electronics Systems, ICCES 2023

作者： Maheswari, B. Subha, S. Tulasi Malini, V.K. Ramanan P, Venkada Rajalakshmi Institute of Technology Department of Computer Science and Engineering Chennai India

ISBN: (纸本)9798350396638

Recognition of handwriting is a critical challenge in artificial intelligence having implications for numerous areas such as document analysis, optical character recognition, including the processing of natural languages. In this study, a novel deep neural network technique for offline handwritten textual data recognition is described. Due to the volume of data available today and the numerous algorithmic improvements ongoing, it is now simpler to construct deep neural networks. Recent advancements in GPU accessibility and numerous cloud-based solutions like AWS or Amazon Web Services, as well as Google's cloud computing platform that offer resources to do so have increased the processing power that is required to simulate neural networks. It has shown great promise to achieve high accuracy in handwriting recognition tasks by incorporating a few deep learning approaches, including Region-based Convolutional neural networks (RCNN), Convolutional neural networks (CNN) along with Recurrent neural networks (RNN) and Long Short-Term Memory (LSTM) networks. The Connectionist Temporal Classification (CTC) method is also frequently used to train deep learning models for sequence identification and loss estimation. © 2023 IEEE.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

A comparative study of Stock Price Prediction Based on BP and LSTM neural network 3

A comparative study of Stock Price Prediction Based on BP an...

引用

3rd International Conference on Artificial Intelligence and Computer Engineering, ICAICE 2022

作者： Huang, Shujia Wang, Ben Hao, Lingbo Si, Zebin School of Information Science and Technology Hangzhou Normal University Hangzhou China

ISBN: (纸本)9781510663473

In recent years, stock price prediction has become a research hotspot. The price of the stock market is unstable, which often rises or falls sharply due to the national policies, which makes it difficult for investors to achieve stable returns in the stock market. With the rapid rise of artificial intelligence, computers have become flexible in dealing with mathematical problems. Therefore, the extraordinary computing power of computers has been used to analyze and predict the trend of the stock market. More and more computer professionals began to enter the financial market and use neural network to study the trend of the stock market. This paper uses BP neural network and LSTM neural network to learn and predict the stock data of Shanghai Composite Index from January 2012 to June 2022. LSTM is a kind of RNN, but it is superior to other neural networks. It can effectively deal with data forgetting and gradient explosion problems, and bring reliability to the prediction results of the model. The two models are evaluated by analyzing MAE, MSE and the time required for model training. The results show that LSTM model can not only learn longer time span than BP model, but also better than BP model in MAE and MSE indexes, which provides some reference and guidance for the prediction of medium and long-term stocks. © 2023 SPIE.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Artificial neural network Training on an optical Processor via Direct Feedback Alignment

Artificial Neural Network Training on an Optical Processor v...

引用

2023 Conference on Lasers and Electro-Optics Europe and European Quantum Electronics Conference, CLEO/Europe-EQEC 2023

作者： Müller, Kilian Launay, Julien Poli, Iacopo Filipovich, Matthew Capelli, Alessandro Hesslow, Daniel Carron, Igor Daudet, Laurent Krzakala, Florent Gigan, Sylvain LightOn 3-5 Impasse Reille Paris75014 France Information Learning and Physics lab. LausanneCH-1015 Switzerland Laboratoire Kastler Brossel ENS-Université PSL CNRS Sorbonne Université Collège de France 24 Rue Lhomond ParisF-75005 France

ISBN: (纸本)9798350345995

Artificial neural networks (ANN) are habitually trained via the back-propagation (BP) algorithm. This approach has been extremely successful: Current models like GPT-3 have O(1011) parameters, are trained on O(1011) words and produce awe-inspiring results. However, there are good reasons to look for alternative training methods: With current algorithms and hardware constraints sometimes only half the available computing power is actually used. This is due to a complicated interplay between the size of the ANN, the available memory, throughput limitations of interconnects, the architecture of the network of computers, and the training algorithm. Training a model like the aforementioned GPT-3 takes months and costs millions. A different training paradigm, which could make clever use of specialized hardware, may train large ANNs more efficiently. © 2023 IEEE.

关键词： neural networks

来源：评论

学校读者我要写书评

暂无评论

Energy-Constrained Model Pruning for Efficient In-Orbit Object Detection in optical Remote Sensing Images 1

引用

7th International Conference on Space Information network (SINC)

作者： Qiu, Shaohua Chen, Du Xu, Xinghua Liu, Jia Naval Univ Engn Natl Key Lab Electromagnet Energy Wuhan 430033 Peoples R China China Univ Geosci Sch Comp Sci Wuhan 430074 Peoples R China East Lake Lab Wuhan 430202 Peoples R China

ISBN: (数字)9789819715688

ISBN: (纸本)9789819715671;9789819715688

Efficient object detection from optical remote sensing (RS) images has always been an important interpretation task for in-orbit RS applications. In recent years, convolutional neural networks have been widely used for object detection with significantly improved detection accuracy. However, the large detection models pose great challenges for the computing, memory and energy supply of resource-constrained in-orbit platforms. In this paper, we propose an efficient in-orbit object detection method with low memory, computation and energy requirements. The proposed method first integrates the compact modules of GhostNet into the detector and further performs the L1-norm based filter pruning to significantly reduce model size and computational complexity. Besides, we propose to use energy as a key metric in filter pruning, and present a novel energy-guided layer-wise pruning rate estimation method so as to achieve energy-efficient object detection. Comprehensive experiments have shown the effectiveness of the proposed method in terms of model size, computational complexity, latency and energy consumption, while maintaining comparable detection accuracy.

关键词： In-Orbit Object Detection optical Remote Sensing Images Constrained Resources Lightweight CNN Filter Pruning

来源：评论

学校读者我要写书评

暂无评论

A preliminary study of functional reconstruction of paralysed hand based on Mediapipe and deep learning for motor intention recognition 6

A preliminary study of functional reconstruction of paralyse...

引用

6th International Conference on Image, Video Processing, and Artificial Intelligence, IVPAI 2024

作者： Ma, Jing Ma, Junhai Wang, Chengzhen He, Weihua Bao, Xueliang Li, Yi School of Information Engineering Ningxia University Ningxia750021 China Key laboratory of artificial intelligence and information security for channeling computing resources from the east to the west Ningxia750021 China

ISBN: (纸本)9781510681781

This paper aims to assess the reliability of a gesture recognition method combining Mediapipe and lightweight deep learning networks as a potential replacement for the EMG-based gesture recognition method in EMG-controlled FES systems. In this paper, the Mediapipe framework was employed to construct a dataset, wherein the 3D coordinates and left/right hand identifiers of 21 key points on the hand were extracted from preprocessed images. The data in the dataset were classified into two categories of data that contained and did not contain left/right hand labelling. A lightweight Dense neural network (DNN) and a lightweight Convolution neural network (CNN) model were constructed and initially trained and tested using the first type of data in the dataset. Subsequently, the lightweight DNN was trained and tested using the second type of data in the dataset, in accordance with the results. The results demonstrated that the lightweight DNN and CNN have a 99.91% and 100% recognition rate, respectively, on the test set of the first type of data. Additionally, the DNN has a 100% recognition rate on the test set of the second type of data. It is noteworthy that the DNN network has a smaller number of model parameters compared to the CNN network. Furthermore, the DNN model demonstrated superior action recognition precision, F1 score, and recall on the test set of the second type of data in comparison to the DNN and CNN models on the first type of data. The hand labelling was beneficial in enhancing the evaluation metrics. © 2024 SPIE.

关键词： Gesture recognition

来源：评论

学校读者我要写书评

暂无评论

2023 8th International Conference on Fog and Mobile Edge computing, FMEC 2023

2023 8th International Conference on Fog and Mobile Edge Com...

引用

8th IEEE International Conference on Fog and Mobile Edge computing, FMEC 2023

ISBN: (纸本)9798350316971

The proceedings contain 39 papers. The topics discussed include: modelling and simulation for detecting vulnerabilities and security threats of smart contracts using machine learning;zero-touch privacy preserving provisioning in an edge-, fog-, and cloud environment;optical character recognition using optimized convolutional networks;botnet detection using label propagation and batch K-means clustering for securing IoT networks;an overview of using of artificial intelligence in enhancing security and privacy in mobile social networks;a federated learning benchmark on tabular data: comparing tree-based models and neural networks;robust-DSN - performance and fault tolerance of a distributed storage network;and trusted federated learning framework for attack detection in Edge industrial Internet of Things.

关键词：

来源：评论

学校读者我要写书评

暂无评论

neural distortion fields for spatial calibration of wide field-of-view near-eye displays

引用

OPTICS EXPRESS 2022年第22期30卷 40628-40644页

作者： Hiroi, Yuichi Someya, Kiyosato Itoh, Yuta Univ Tokyo 7-3-1 HongoBunkyo Ku Tokyo 1138654 Japan Tokyo Inst Technol 4259 NagatsutachoMidoriku Yokohama Kanagawa 2260026 Japan

We propose a spatial calibration method for wide field-of-view (FoV) near-eye displays (NEDs) with complex image distortions. Image distortions in NEDs can destroy the reality of the virtual object and cause sickness. To achieve distortion-free images in NEDs, it is necessary to establish a pixel-by-pixel correspondence between the viewpoint and the displayed image. Designing compact and wide-FoV NEDs requires complex optical designs. In such designs, the displayed images are subject to gaze-contingent, non-linear geometric distortions, which explicit geometric models can be difficult to represent or computationally intensive to optimize. To solve these problems, we propose neural distortion field (NDF), a fully-connected deep neural network that implicitly represents display surfaces complexly distorted in spaces. NDF takes spatial position and gaze direction as input and outputs the display pixel coordinate and its intensity as perceived in the input gaze direction. We synthesize the distortion map from a novel viewpoint by querying points on the ray from the viewpoint and computing a weighted sum to project output display coordinates into an image. Experiments showed that NDF calibrates an augmented reality NED with 90 degrees FoV with about 3.23 pixel (5.8 arcmin) median error using only 8 training viewpoints. Additionally, we confirmed that NDF calibrates more accurately than the non-linear polynomial fitting, especially around the center of the FoV.(c) 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement

关键词： Geometric optics Image metrics Image quality Lens design optical aberrations optical design

来源：评论

学校读者我要写书评

暂无评论

A bio-inspired design of real-time control learning algorithm in model-free drone manipulation

A bio-inspired design of real-time control learning algorith...

引用

2024 International Conference on Optics, Electronics, and Communication Engineering, OECE 2024

作者： Wang, Jiaguo Li, Wenheng Pei, Yang Northwestern Polytechnical University Shaanxi Xi’an710072 China AVIC Xi’an Aeronautics Computing Technique Research Institute Shaanxi Xi’an710068 China

ISBN: (纸本)9781510685437

In recent years, neural-network-based adaptive dynamic models are commonly used to estimate and control flight dynamics for drones and multi-copters. However, most of them simply use networks to optimize few parameters in control policy. There are still large improvements on the structural layout for robust model-free control applications. Therefore, to achieve a similar intelligence of human is still challenging due to their difference in basic mechanisms and difficulty in network modeling. In this paper, we design a control learning algorithm which combines reinforcement learning with neural networks simplified from human cerebellar motor learning model. The algorithm learns parameters by statistically measuring the performance and analyzing the input-output relationship on real-time episodes. In local linear systems, parameters are learned with respect to a spatial function of environment state and subjective expectation. Compared with other methods using static models, the most obvious advantage of this algorithm is that it can learn complex dynamics of alternative degrees of freedom while the dynamics are difficult to be formulated by equation set. Besides, the algorithm is suitable for individual systems, without prior knowledge about system geometry, centre of gravity as well as installation error, since it learns the dynamic effects directly relevant to the optimal guidance and control behavior in unknown or partially known environments instead. Experiment verifies the algorithm in a practical way. In the experiment, the algorithm is implemented to a quad-copter and it can learn the flight control policy from zero-state without any prior knowledge. The flight quality is tested to be equal to accurate control model at outdoor flight experiences. By repeated experiment, the algorithm is demonstrated to have good robustness to control different physical models and the potential to explore alternative dimensionality. © 2024 SPIE.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

Photonic Convolution Engine based on Phase-Change Materials and Stochastic computing 26

Photonic Convolution Engine based on Phase-Change Materials ...

引用

26th IEEE-Computer-Society Annual Symposium on VLSI (ISVLSI)

作者： Cardoso, Raphael Zrounba, Clement Abdalla, Mohab Jimenez, Paul Gomes, Mauricio Charbonnier, Benoit Pavanello, Fabio O'Connor, Ian Le Beux, Sebastien Univ Lyon Univ Claude Bernard Lyon 1 Ecole Cent Lyon INSA LyonCNRSCPE LyonINLUMR5270 Ecully F-69134 Lyon France RMIT Univ Sch Engn Melbourne Vic 3000 Australia Univ Grenoble Alpes CEA LETI F-38000 Grenoble France Concordia Univ Montreal Dept Elect & Comp Engn Montreal PQ Canada

ISBN: (纸本)9798350327694

The last wave of AI developments sparked a global surge in computing resources allocated to neural network models. Even though such models solve complex problems, their mathematical foundations are simple, with the multiply-accumulate (MAC) operation standing out as one of the most important. However, improvements in traditional CMOS technologies fail to match the ever-increasing performance requirements of AI applications, therefore new technologies, as well as disruptive computing architectures must be explored. In this paper, we propose a novel in-memory implementation of a MAC operator based on stochastic computing and optical phase-change memories (oPCMs), leveraging their proven non-volatility and multilevel capabilities to achieve convolution. We show that resorting to the stochastic computing paradigm allows one to exploit the dynamic mechanisms of oPCMs to naturally compute and store MAC results with less noise sensitivity. Under similar conditions, we demonstrate an improvement of up to 64x and 10x in the applications that we evaluated.

关键词： photonic computing phase-change memories stochastic computing in-memory computing

来源：评论

学校读者我要写书评

暂无评论

Advancing optical Character Recognition for Handwritten Text: Enhancing Efficiency and Streamlining Document Management 14

Advancing Optical Character Recognition for Handwritten Text...

引用

14th International Conference on computing Communication and networking Technologies, ICCCNT 2023

作者： Benedict Vinusha, V. Indhuja, V. Varshitha Reddy, M. Nikhitha, N. Reddy, Naga Venkata Siva R.M.K. Engineering College Department of Artificial Intelligence and Data Science Kavaraipettai601206 India

ISBN: (纸本)9798350335095

This research paper explores the challenge of converting handwritten text into editable digital text, in both native and English languages. We propose optical character recognition (OCR) as a solution, leveraging algorithms and machine learning models to recognize and convert text into various digital formats. The paper highlights the significance of OCR in digitizing paper documents and streamlining document management across diverse industries. Specifically, the study focuses on developing OCR software capable of identifying regional languages, providing a comprehensive assessment of its precision and effectiveness. The findings demonstrate how OCR technology can significantly improve efficiency and cost-effectiveness for businesses dealing with large volumes of paper documents. Our approach involves using image inputs, analyzing them with Keras and OpenCV, and implementing Convolutional neural network (CNN) techniques and Natural Language Processing (NLP) libraries. © 2023 IEEE.

关键词： Cost effectiveness

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：