检索结果-内蒙古大学图书馆

Convergence of gradient descent-ascent analyzed as a Newtonian dynamical system with dissipation

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Seung, H. Sebastian Howard Hughes Medical Institute Dept. of Brain & Cog. Sci. Dept. of Physics MIT Neuroscience Institute and Computer Science Department Princeton University

A dynamical system is defined in terms of the gradient of a payoff function. Dynamical variables are of two types, ascent and descent. The ascent variables move in the direction of the gradient, while the descent variables move in the opposite direction. Dynamical systems of this form or very similar forms have been studied in diverse fields such as game theory, optimization, neural networks, and population biology. Gradient descent-ascent is approximated as a Newtonian dynamical system that conserves total energy, defined as the sum of the kinetic energy and a potential energy that is proportional to the payoff function. The error of the approximation is a residual force that violates energy conservation. If the residual force is purely dissipative, then the energy serves as a Lyapunov function, and convergence of bounded trajectories to steady states is guaranteed. A previous convergence theorem due to Kose and Uzawa required the payoff function to be convex in the descent variables, and concave in the ascent variables. Here the assumption is relaxed, so that the payoff function need only be globally "less convex" or "more concave" in the ascent variables than in the descent variables. Such relative convexity conditions allow the existence of multiple steady states, unlike the convex-concave assumption. When combined with sufficient conditions that imply the existence of a minimax equilibrium, boundedness of trajectories is also assured.137N40, 90C26 Copyright © 2019, The Authors. All rights reserved.

关键词： Dynamical systems

Employing best input SVR robust lost function with nature-inspired metaheuristics in wind speed energy forecasting

学校读者我要写书评

暂无评论

IAENG International Journal of Computer Science

IAENG International Journal of computer science 2020年第3期47卷 572-584页

作者： Caraka, Rezzy Eko Chen, Rung Ching Bakar, Sakhinah Abu Tahmid, Muhammad Toharudin, Toni Pardamean, Bens Huang, Su-Wen College of Informatics Chaoyang University of Technology Taichung City41349 Taiwan School of Mathematical Sciences The National University of Malaysia Bandar Baru Bangi Selangor43600 Malaysia Bioinformatics and Data Science Research Center Bina Nusantara University 11480 Indonesia Manado North Sulawesi91594 Indonesia Department of Statistics Padjadjaran University Jatinagor west java45363 Indonesia Bioinformatics and Data Science Research Center Bina Nusantara University and Computer Science Dept. BINUS Graduate Program-Master of Computer Science Program Bina Nusantara University Jakarta11480 Indonesia Taichung Veterans General Hospital Taiwan

Wind power has been experiencing a quick improvement. Without a doubt, wind is a variable asset that is hard to forecast. For instance, traditionally time series, extra holds are distributed to deal with this uncertainty. This paper presents a comparison of the performance of various Support Vector Regression (SVR) applied to short-term wind power forecasting. The analogy with BORUTA and multivariate adaptive regression splines (MARS) as judge best input and employ genetic algorithm and particle swarm optimization to find best parameter in Support Vector Regression with robust lost function. We measure the accuracy of this models by Symmetric means absolute percentage error (sMAPE) and we get the best model BORUTA-SVR-PSO with sMAPE 2.07155%. Moreover, we measure the energy conversion using Feedback Linearization Control (FLC). © International Association of Engineers.

关键词： Genetic algorithms

Implementation of real-time scanner Java language text with mobile vision Android based 1

学校读者我要写书评

暂无评论

Implementation of real-time scanner Java language text with ...

1st International Conference on Information and Communications Technology, ICOIACT 2018

作者： Nurzam, Fariz Dzulfiqar Luthfi, Emha Taufiq Dept. of Informatics International Program Faculty of Computer Science University of AMIKOM Yogyakarta Yogyakarta Indonesia

ISBN: (纸本)9781538609545

Currently technological research in the translating has penetrated to image processing. One of the role of image processing is optical character recognition technology on various platform, an example on Android OS. However, the current research on image processing of translating with the native regional language of Indonesia still rarely. In this research, writers performs the implementation of Latin text translate from Bahasa Indonesia into Javanese text with Google Mobile Vision in real-time, also vice versa with Android mobile based application. Which is Google Mobile Vision is one of the open API optical character recognition library with pattern recognition method. With the design of client server architecture, writers develop Android application as client and Web Service that can process REST API and interact with the database. The output of this research is an application for translate that can scan words in Javanese into Bahasa Indonesia, or vice versa in frame of an Android application. By this research, it expect to be the benchmark for image processing research for native regional language of Indonesia. © 2018 IEEE.

关键词： Optical character recognition

Comparing CART and C5.0 algorithm performance of human development index 3

学校读者我要写书评

暂无评论

Comparing CART and C5.0 algorithm performance of human devel...

3rd International Conference on Informatics and Computing, ICIC 2018

作者： Assiroj, Priati Warnars, Harco Leslie Hendric Spits Fauzi, Ahmad Dept. of Information System Faculty of Technology and Computer Science Buana Perjuangan University Karawang Indonesia Computer Science Department BINUS Graduate Program Doctor of Computer Science Bina Nusantara University Jakarta11480 Indonesia Dept. of Computer Science Faculty of Technology and Computer Science Buana Perjuangan University Karawang Indonesia

ISBN: (纸本)9781538669204

Human Development Index (HDI) is a reference to how far the administration can develops their people. The development of a territory is related to a prosperity of the people. This research is focused on Human Development Index in West Java, a Province of Indonesia. Research identifies variables which are influence and very important to HDI and compares the model of CART and C5.0 algorithm based on HDI concept that has developed by United Nation (UN). Research uses 14 variables which are commonly used for measuring HDI and found that there are two variables that most influencing to HDI, MYS-index and EDU-Index. By knowing the MYS and EDU-Index research can predict and classify each territory to its HDI. Research could be conducted to predict HDI in the other territory. The results evaluated by measuring the accuracy and confusion matrix, 91.35 for CART algorithm and 90.80 for C5.0 algorithm. © 2018 IEEE.

关键词： Algorithm Classification Comparison EDU-Index Human Development Index (HDI) MYS-Index Territory

A therapy game for elementary students with ADHD

学校读者我要写书评

暂无评论

AIP Conference Proceedings 2023年第1期2508卷

作者： Raymond Bahana Edi Abdurachman Ford L. Gaol Tjhin Wiguna Fonny Dameaty Hutagalung Bayu Prakoso Dirgantoro Eko Nugroho 1Computer Science Department BINUS Graduate Program Doctor of Computer Science Bina Nusantara University Jakarta Indonesia 2Faculty of Medicine Universitas Indonesia Depok Indonesia 3Department of Educational Psychology and Counselling Faculty of Education University of Malaya Kuala Lumpur Malaysia 4Graphic Design and New Media Prog. Visual Communication Design Dept. BINUS Northumbria School of Design Bina Nusantara University Jakarta Indonesia 5Kummara Bandung Indonesia

Children with Attention Deficit Hyperactivity Disorder usually have problem to express themselves. Using game as a therapy, they could reduce the problem. A recent study has revealed video games might be an alternative therapy for children with ADHD. This research is to analyze the data that was collected from a prototype game that has been developed and created. The game is created using Construct2. To collect data from the game, back-end web application was created using PHP and MySQL. Several data that have been used in this research paper are score, incorrect house, and incorrect fruit. Ten children with ADHD are involved in this testing, all children are elementary students and never receive any treatment for their ADHD. Before and after the testing, they have been tested by fMRI. The testing consists of 4 weeks of training sessions of about 20-30 minutes almost every day except Sunday or public holiday. Every session contains at least three tutorial or training tasks and four game tasks in a fixed order. In conclusion, the result showed the scores were better when the children with ADHD play with visual instruction rather than with audio instruction.

关键词：

Uncovering the Folding Landscape of RNA Secondary Structure with Deep Graph Embeddings

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Castro, Egbert Benz, Andrew Tong, Alexander Wolf, Guy Krishnaswamy, Smita Comp. Biol. and Bioinf. Program Yale University New HavenCT United States Dept. of Mathematics Yale University New HavenCT United States Dept. of Computer Science Yale University New HavenCT United States Dept. of Math Stat. Univ. de Montréal Mila MontrealQC Canada Depts. of Genetics & Comp. Sci. Yale University New HavenCT United States

Biomolecular graph analysis has recently gained much attention in the emerging field of geometric deep learning. Here we focus on organizing biomolecular graphs in ways that expose meaningful relations and variations between them. We propose a geometric scattering autoencoder (GSAE) network for learning such graph embeddings. Our embedding network first extracts rich graph features using the recently proposed geometric scattering transform. Then, it leverages a semi-supervised variational autoencoder to extract a low-dimensional embedding that retains the information in these features that enable prediction of molecular properties as well as characterize graphs. We show that GSAE organizes RNA graphs both by structure and energy, accurately reflecting bistable RNA structures. Also, the model is generative and can sample new folding trajectories. Copyright © 2020, The Authors. All rights reserved.

关键词： RNA

Incomplete nutation diffusion imaging: An ultrafast, single-scan approach for diffusion mapping

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Ianuş, Andrada Shemesh, Noam Champalimaud Neuroscience Programme Champalimaud Centre for the Unknown Lisbon Portugal Centre for Medical Image Computing Dept. of Computer Science University College London London United Kingdom

Purpose: Diffusion Magnetic Resonance Imaging (dMRI) is confounded by its long acquisition duration, thereby thwarting the detection of rapid microstructural changes, especially when diffusivity variations are accompanied by rapid changes in T2. The purpose of the present study is to accelerate dMRI to a single scan acquisition, and to enable a more accurate estimation of diffusivity as function of time. Methods: A general methodology termed Incomplete Initial Nutation Diffusion Imaging (INDI) capturing two diffusion contrasts in a single shot, is presented. INDI creates a longitudinal magnetization reservoir that facilitates the successive acquisition of two images separated by only a few milliseconds. INDI's theory is presented, followed by proof-of-concept ex- and in-vivo experiments at 16.4 T and 9.4 T. Results: Mean diffusivities (MDs) extracted from INDI were comparable with Diffusion Tensor Imaging (DTI) and the two-shot IDE in the water phantom. As expected in the brain tissues, DTI provided lower MD than UF-IDE and IDE, but IDE and UF-IDE were in excellent agreement. Simulations are presented for identifying the regimes where INDI is most beneficial. Conclusions: INDI accelerates dMRI acquisition to single-shot mode, which can be of great importance for mapping dynamic microstructural properties in-vivo without T2 bias. Copyright © 2019, The Authors. All rights reserved.

关键词： Magnetic resonance imaging

Learning and segmenting dense voxel embeddings for 3D neuron reconstruction

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Lee, Kisuk Lu, Ran Luther, Kyle Seung, H. Sebastian The Brain & Cog. Sci. Dept. Massachusetts Institute of Technology CambridgeMA02139 United States The Neuroscience Institute Princeton University PrincetonNJ08544 United States The Dept. of Physics Princeton University PrincetonNJ08544 United States The Dept. of Computer Science The Neuroscience Institute Princeton University PrincetonNJ08544 United States

We show dense voxel embeddings learned via deep metric learning can be employed to produce a highly accurate segmentation of neurons from 3D electron microscopy images. A "metric graph" on a set of edges between voxels is constructed from the dense voxel embeddings generated by a convolutional network. Partitioning the metric graph with long-range edges as repulsive constraints yields an initial segmentation with high precision, with substantial accuracy gain for very thin objects. The convolutional embedding net is reused without any modification to agglomerate the systematic splits caused by complex "self-contact" motifs. Our proposed method achieves state-of-the-art accuracy on the challenging problem of 3D neuron reconstruction from the brain images acquired by serial section electron microscopy. Our alternative, object-centered representation could be more generally useful for other computational tasks in automated neural circuit reconstruction. Copyright © 2019, The Authors. All rights reserved.

关键词： Electron microscopy

Automatic detection and compression for passive acoustic monitoring of the African forest elephant

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Bjorck, Johan Rappazzo, Brendan H. Chen, Di Bernstein, Richard Wrege, Peter H. Gomes, Carla P. Dept. of Computer Science Cornell University IthacaNY United States Bioacoustics Research Program Cornell Lab of Ornithology IthacaNY United States

In this work, we consider applying machine learning to the analysis and compression of audio signals in the context of monitoring elephants in sub-Saharan Africa. Earths biodiversity is increasingly under threat by sources of anthropogenic change (e.g. resource extraction, land use change, and climate change) and surveying animal populations is critical for developing conservation strategies. However, manually monitoring tropical forests or deep oceans is intractable. For species that communicate acoustically, researchers have argued for placing audio recorders in the habitats as a cost-effective and non-invasive method, a strategy known as passive acoustic monitoring (PAM). In collaboration with conservation efforts, we construct a large labeled dataset of passive acoustic recordings of the African Forest Elephant via crowdsourcing, compromising thousands of hours of recordings in the wild. Using state-of-the-art techniques in artificial intelligence we improve upon previously proposed methods for passive acoustic monitoring for classification and segmentation. In real-time detection of elephant calls, network bandwidth quickly becomes a bottleneck and efficient ways to compress the data are needed. Most audio compression schemes are aimed at human listeners and are unsuitable for low-frequency elephant calls. To remedy this, we provide a novel end-to-end differentiable method for compression of audio signals that can be adapted to acoustic monitoring of any species and dramatically improves over nive coding strategies. Copyright © 2019, The Authors. All rights reserved.

关键词： Cost effectiveness