检索结果-南通市图书馆

Feature Representations Using the Reflected Rectified Linear Unit(RReLU) Activation

Big Data Mining and Analytics 2020年第2期3卷 102-120页

作者： chaity banerjee Tathagata Mukherjee Eduardo Pasiliao Jr. Department of Idustrial&Systems Engineering University of Central FloridaFloridaOrlandoFL 32816-2368USA Department of Computer Science University of Alabama in HuntsvilleHuntsvilleHuntsvilleAL 35806USA Air Force Research Labs United States Air ForceEglin Air Force BaseShalimarFL 32579USA

Deep Neural Networks(DNNs)have become the tool of choice for machine learning practitioners *** important aspect of designing a neural network is the choice of the activation function to be used at the neurons of the different *** this work,we introduce a four-output activation function called the Reflected Rectified Linear Unit(RRe LU)activation which considers both a feature and its negation during *** activation function is"sparse",in that only two of the four possible outputs are active at a given *** test our activation function on the standard MNIST and CIFAR-10 datasets,which are classification problems,as well as on a novel Computational Fluid Dynamics(CFD)dataset which is posed as a regression *** the baseline network for the MNIST dataset,having two hidden layers,our activation function improves the validation accuracy from 0.09 to 0.97 compared to the well-known Re LU *** the CIFAR-10 dataset,we use a deep baseline network that achieves 0.78 validation accuracy with 20 epochs but overfits the *** the RRe LU activation,we can achieve the same accuracy without overfitting the *** the CFD dataset,we show that the RRe LU activation can reduce the number of epochs from 100(using Re LU)to 10 while obtaining the same levels of performance.

关键词： deep learning feature space approximations multi-output activations Rectified Linear Unit(ReLU)

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Deep Feature Learning for Intrinsic Signature Based Camera Discrimination

引用

Big Data Mining and Analytics 2022年第3期5卷 206-227页

作者： chaity banerjee Tharun Kumar Doppalapudi Eduardo Pasiliao Jr. Tathagata Mukherjee Department of Idustrial&Systems Engineering University of Central FloridaOrlandoFL 32816USA Department of Computer Science University of Alabama in HuntsvilleHuntsvilleAL 35806USA Air Force Research Labs United States Air ForceEglin Air Force BaseShalimarFL 32579USA

In this paper we consider the problem of“end-to-end”digital camera identification by considering sequence of images obtained from the *** problem of digital camera identification is harder than the problem of identifying its analog counterpart since the process of analog to digital conversion smooths out the intrinsic noise in the analog *** it is known that identifying a digital camera is possible by analyzing the camera’s intrinsic sensor artifacts that are introduced into the images/videos during the process of photo/video *** is known that such methods are computationally intensive requiring expensive pre-processing *** this paper we propose an end-to-end deep feature learning framework for identifying cameras using images obtained from *** conduct experiments using three custom datasets:the first containing two cameras in an indoor environment where each camera may observe different scenes having no overlapping features,the second containing images from four cameras in an outdoor setting but where each camera observes scenes having overlapping features and the third containing images from two cameras observing the same checkerboard pattern in an indoor *** results show that it is possible to capture the intrinsic hardware signature of the cameras using deep feature representations in an end-to-end *** deep feature maps can in turn be used to disambiguate the cameras from each *** system is end-to-end,requires no complicated pre-processing steps and the trained model is computationally efficient during testing,paving a way to have near instantaneous decisions for the problem of digital camera identification in production *** we present comparisons against the current state-of-the-art in digital camera identification which clearly establishes the superiority of the end-to-end solution.

关键词： deep learning visual signatures camera identification convolutional neural networks deep feature learning

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Optimal harvesting of a prey-predator model with variable carrying capacity

引用

International Journal of Biomathematics 2017年第5期10卷 185-208页

作者： chaity Ganguli T. K. Kart P. K. Mondal Department of Mathematics Indian Institute of Engineering Science and Technology Shibpur Howrah 711103 West Bengal India Department of Mathematics Shibpur Sri Ramkrishna Vidyalaya Howrah 711102 West Bengal India

This work deals with a prey-predator model in an environment where the carrying capacities are assumed to be variable with time and one species feeds upon the other. Independent harvesting efforts are applied in either species and asymmetrical intraguild predation occurs. A common resource is consumed by two competing species and at the same time predator also consumes the prey. At first we discuss the model under constant carrying capacity and make the conclusion that no limit cycle exists in this case. Then we discuss the model without intraspecific competition. Our main concern is to cover the above mentioned two cases together, i.e. the model with variable carrying capacity and intraspecific competition. We determine the steady states and examine the dynamical behavior. We also analyze the local and global stability of the interior equilibrium by Routh-Hurwitz criterion and a suitable Lyapunov function respectively. A Hopf bifur- cation occurs with respect to a parameter which is the ratio of predator＇s and prey＇s intrinsic growth rate. The possibility of bionomic equilibrium has been considered. The optimal harvest policy is formulated and solved with Pontryagin＇s maximum principle. Some numerical simulations are given to explain most of the analytical results.

关键词： Variable carrying capacity intraguild predation Hopf bifurcation optimal harvesting.

来源：

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Photonic Crystal Fiber Based Sensor for Detecting Binary Liquid Mixture

引用

Optics and Photonics Journal 2017年第11期7卷 221-234页

作者： Md. Faizul Huq Arif Md. Jaminul Haque Biddut Md. Shaiful Islam Babu H M Mostafizur Rahman Md. Mahbubur Rahman Busrat Jahan Moshrefa Sultana chaity Shah Mostafa Khaled Institute of Information Technology (IIT) University of Dhaka Dhaka Bangladesh Automation and Robotics Faculty of Electrical Engineering and Information Technology (EEIT) TU Dortmund University Dortmund Germany Department of Computer Science and Engineering World University of Bangladesh Dhaka Bangladesh School of Information Technology BAC International Study Centre Dhaka Bangladesh Department of Computer Science and Engineering Feni University Feni Bangladesh Institute of Information Technology (IIT) Jahangirnagar University Dhaka Bangladesh

Detection of liquid mixture of different volume ratio is very important in industrial purposes. The paper reports a sensing mechanism of binary liquid mixture for different volume fraction, based on the measurement of refractive index of the mixture. Here, a highly sensitive liquid filled core Photonic Crystal Fiber structure has been proposed to detect liquid mixture solution. Numerical investigation of the proposed structure is carried out by employing full vectorial Finite Element Method (FEM).

关键词： Binary Liquid Mixture Photonic Crystal Fiber Refractive Index Sensitivity Confinement Loss

来源：

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

欢迎您,

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

在线全文

在线全文

在线全文

在线全文

请选择保存的检索档案：

请选择收藏分类：

通借通还

欢迎您,

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

在线全文

在线全文

在线全文

在线全文

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：