• Geophysics and Geophysical Exploration
• /
• v.25 no.4
• /
• pp.242-251
• /
• 2022

Automatic differentiation automatically calculates the derivatives of a function using the chain rule once the forward operation of a function is defined. Given the recent development of computing libraries that support automatic differentiation, many researchers have adopted automatic differentiation techniques to solve geophysical inverse problems. We analyzed the advantages, disadvantages, and performances of automatic differentiation techniques using the gradient calculations of seismic full waveform inversion objective functions. The gradients of objective functions can be expressed as multiplications of the derivatives of the model parameters, wavefields, and objective functions using the chain rule. Using numerical examples, we demonstrated the speed of analytic differentiation and the convenience of complex gradient calculations for automatic differentiation. We calculated derivatives of model parameters and objective functions using automatic differentiation and derivatives of wavefields using analytic differentiation.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.36 no.7
• /
• pp.634-641
• /
• 2008

Automatic Differentiation(AD) is a tool for generating sensitivities, such as gradient or Jacobian, automatically. AD tools provide mathematically exact sensitivities for the given source code. In this paper applications of automatic differentiation are studied. Derivative codes are generated with AD tools for structural analysis code and flow analysis code. How to apply AD tools is explained and the accuracy of sensitivities is compared with the finite difference. Sensitivities of generated derivative code accord well with finite difference, but the calculation time of derivative code increases. It was found that the calculation time can be decreased by additional modification of derivative code.

• Journal of Korean Society of Steel Construction
• /
• v.12 no.5 s.48
• /
• pp.569-579
• /
• 2000

An improved multi-level (IML) optimization algorithm using automatic differentiation (AD) of framed structures is proposed in this paper. For the efficiency of the proposed algorithm, multi-level optimization techniques using a decomposition method that separates both system-level and element-level optimizations, that utilizes and an artificial constraint deletion technique, are incorporated in the algorithm. And also to save the numerical efforts, an efficient reanalysis technique through approximated structural responses such as moments and frequencies with respect to intermediate variables is proposed in the paper. Sensitivity analysis of dynamic structural response is executed by AD that is a powerful technique for computing complex or implicit derivatives accurately and efficiently with minimal human effort. The efficiency and robustness of the IML algorithm, compared with a plain multi-level (PML) algorithm, is successfully demonstrated in the numerical examples.

• Journal of the Korean Society for information Management
• /
• v.24 no.1 s.63
• /
• pp.251-271
• /
• 2007

This paper studies the problem of classifying documents with labeled and unlabeled learning data, especially with regards to using document similarity features. The problem of using unlabeled data is practically important because in many information systems obtaining training labels is expensive, while large quantities of unlabeled documents are readily available. There are two steps In general semi-supervised learning algorithm. First, it trains a classifier using the available labeled documents, and classifies the unlabeled documents. Then, it trains a new classifier using all the training documents which were labeled either manually or automatically. We suggested two types of semi-supervised learning algorithm with regards to using document similarity features. The one is one step semi-supervised learning which is using unlabeled documents only to generate document similarity features. And the other is two step semi-supervised learning which is using unlabeled documents as learning examples as well as similarity features. Experimental results, obtained using support vector machines and naive Bayes classifier, show that we can get improved performance with small labeled and large unlabeled documents then the performance of supervised learning which uses labeled-only data. When considering the efficiency of a classifier system, the one step semi-supervised learning algorithm which is suggested in this study could be a good solution for improving classification performance with unlabeled documents.

• 한국전산유체공학회:학술대회논문집
• /
• 2004.10a
• /
• pp.181-186
• /
• 2004

In gradient based optimization methods, the finite differencing which uses small perturbations in the design variables has been used to calculate the sensitivity. Recently, the automatic differentiation has been widely studied to calculate the function value and the sensitivities simultaneously. In this paper, the applicability of the automatic differentiation In the aerodynamic design optimization is studied. ADIFOR and TAPENADE are used to generate the codes which give the function value and the sensitivities for 2D compressible inviscid flows.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.24 no.7
• /
• pp.842-848
• /
• 2020

Music markets have entered the era of streaming. In order to select and propose music that suits the taste of music consumers, there is an active demand and research on an automatic music genre classification system. We propose a method to improve the accuracy of genre unclassified songs, which was a lack of the previous system, by using a generative adversarial network (GAN) to further develop the automatic voting system for deep learning music genre using Softmax proposed in the previous paper. In the previous study, if the spectrogram of the song was ambiguous to grasp the genre of the song, it was forced to leave it as an unclassified song. In this paper, we proposed a system that increases the accuracy of genre classification of unclassified songs by converting the spectrogram of unclassified songs into an easy-to-read spectrogram using GAN. And the result of the experiment was able to derive an excellent result compared to the existing method.

• ICROS
• /
• v.2 no.1
• /
• pp.72-84
• /
• 1996

• Computational Structural Engineering
• /
• v.35 no.4
• /
• pp.23-26
• /
• 2022

• Computational Structural Engineering
• /
• v.36 no.4
• /
• pp.14-17
• /
• 2023

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.25 no.2
• /
• pp.182-191
• /
• 2001

The paper describes the improvement on concurrent subspace optimization(CSSO) via automatic differentiation. CSSO is an efficient strategy to coupled multidisciplinary design optimization(MDO), wherein the original design problem is non-hierarchically decomposed into a set of smaller, more tractable subspaces. Key elements in CSSO are consisted of global sensitivity equation, subspace optimization, optimum sensitivity analysis, and coordination optimization problem that require frequent use of 1st order derivatives to obtain design sensitivity information. The current version of CSSO adopts automatic differentiation scheme to provide a robust sensitivity solution. Automatic differentiation has numerical effectiveness over finite difference schemes tat require the perturbed finite step size in design variable. ADIFOR(Automatic Differentiation In FORtran) is employed to evaluate sensitivities in the present work. The use of exact function derivatives facilitates to enhance the numerical accuracy during the iterative design process. The paper discusses how much the automatic differentiation based approach contributes design performance, compared with traditional all-in-one(non-decomposed) and finite difference based approaches.