• Proceedings of the KIEE Conference
• /
• 1997.11a
• /
• pp.600-602
• /
• 1997

In this study, a hybrid neural net consisting of an Adaptive LVQ(ALVQ) algorithm and MLP is proposed to perform speaker identification task. ALVQ is a new learning procedure using adaptively feature vector sequence instead of only one feature vector in training codebooks initialized by LBG algorithm and the optimization criterion of this method is consistent with the speaker classification decision rule. ALVQ aims at providing a compressed, geometrically consistent data representation. It is fit to cover irregular data distributions and computes the distance of the input vector sequence from its nodes. On the other hand, MLP aim at a data representation to fit to discriminate patterns belonging to different classes. It has been shown that MLP nets can approximate Bayesian "optimal" classifiers with high precision, and their output values can be related a-posteriori class probabilities. The different characteristics of these neural models make it possible to devise hybrid neural net systems, consisting of classification modules based on these two different philosophies. The proposed method is compared with LBG algorithm, LVQ algorithm and MLP for performance.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.3
• /
• pp.125-132
• /
• 2022

Sentence compression is a natural language processing task that generates concise sentences that preserves the important meaning of the original sentence. For grammatically appropriate sentence compression, early studies utilized human-defined linguistic rules. Furthermore, while the sequence-to-sequence models perform well on various natural language processing tasks, such as machine translation, there have been studies that utilize it for sentence compression. However, for the linguistic rule-based studies, all rules have to be defined by human, and for the sequence-to-sequence model based studies require a large amount of parallel data for model training. In order to address these challenges, Deleter, a sentence compression model that leverages a pre-trained language model BERT, is proposed. Because the Deleter utilizes perplexity based score computed over BERT to compress sentences, any linguistic rules and parallel dataset is not required for sentence compression. However, because Deleter compresses sentences only considering perplexity, it does not compress sentences by reflecting the linguistic information of the words in the sentences. Furthermore, since the dataset used for pre-learning BERT are far from compressed sentences, there is a problem that this can lad to incorrect sentence compression. In order to address these problems, this paper proposes a method to quantify the importance of linguistic information and reflect it in perplexity-based sentence scoring. Furthermore, by fine-tuning BERT with a corpus of news articles that often contain proper nouns and often omit the unnecessary modifiers, we allow BERT to measure the perplexity appropriate for sentence compression. The evaluations on the English and Korean dataset confirm that the sentence compression performance of sentence-scoring based models can be improved by utilizing the proposed method.

• Journal of the Korean Chemical Society
• /
• v.16 no.1
• /
• pp.25-32
• /
• 1972

The effect on the apparent viscosity of 2 wt. % Veegum suspensions of different types of electrolytes and of different electrolyte concentrations was studied. Measurements were made with a Brookfield Synchro-Lectric Viscometer, using no.3 spindle at 30 R.P.M. at $24^{\circ}C$. As electriolyte concentration increased, the apparant viscosity was observed to increase to a maximum and then to decrease. Changes in viscosity were in general agreement with predicted results based on the Hofmeister sequence and the Schulze-Hardy rule. The observed electrolyte effect on the apparent viscosity was discussed in terms of the Verwey-Overbeek theory.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.7
• /
• pp.3397-3413
• /
• 2016

Considering that the layered cellular automata (LCA) are naturally fit for representing image data in various applications, a novel reverse iterative image encryption scheme based on LCA is proposed. Specifically, the plain image is set as the final configuration of an 8-layer CA, and some sequences derived from a random sequence are set as the pre-final configuration, which ensure that the same plain image will never be encrypted in the same way when encrypted many times. Then, this LCA is backward evolved by following some reversible two order rules, which are generated with the aid of a newly defined T-shaped neighborhood. The cipher image is obtained from the recovered initial configuration. Several analyses and experimental results show that the proposed scheme possesses a high security level and executive performance.

• Transactions of Materials Processing
• /
• v.8 no.6
• /
• pp.591-603
• /
• 1999

A computer-aided process planning system for rotationally symmetric deep drawing products has been developed. The application for non-axisymmetric components, however, has been reported yet. Therefore, this study investigates process sequence design in deep drawing process and constructs a computer-aided process planning system for non-axisymmetric motor frame products with elliptical shape. The system developed consists of three modules. The first one os a 3-dimensional modeling module to calculate surface area for non-axisymmetric products. The second one is a blank design module that creates an oval-shaped blank with the identical surface area. The third one is a process planning module based on production rules that play the best important roles in an expert system for manufacturing. The production rules are generated and upgraded by interviewing with field engineers. Especially, drawing coefficient, punch and die radii are considered as main design parameters. The constructed system for elliptical deep drawing products would be very useful to reduce lead time and improve accuracy for production.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.8 no.5
• /
• pp.243-248
• /
• 2013

Reliability evaluation is important task in embedded system. It can avoid potential failures and manage the vulnerable components of embedded system effectively. Dynamic fault tree analysis is one of the reliability evaluation methods. It can represent dynamic characteristics of a system such as fault & error recovery, sequence-dependent failures. In this paper, the steering system, which is embedded system in vehicles, is represented using dynamic fault tree. We evaluate the steering system using approximation algorithm based on Simpson's rule. A set of simulation results shows that proposed method overcomes the low accuracy of classic approximation method without requiring no excessive calculation time of the Markov chain method.

• Journal of the Architectural Institute of Korea Planning & Design
• /
• v.35 no.4
• /
• pp.15-24
• /
• 2019

Currently, museums and architectural heritage provide augmented user experiences by incorporating various media technologies. They still, however, suffer from the limitation of entertainment-based and the provision of location-based simple and repetitive contents. In addition, while acting as a key medium of experience for architectural heritage, the concept of space is not properly reflected in current services. The purpose of this study is to design user space experience considering such characteristics of architectural heritage. The spatial experience content and content production platform are defined. This software platform creates content that enhances the experience of the place by giving a context-based digital data associated with space and objects. The spatial experience content is designed as a series of experience sequences. The composition of the sequence borrows the method of film and narrative which segment and connect consecutive experiences on a scene basis considering user's detailed spatial experience. Therefore, content components can be combined and reproduced in various types. Augmented contents were extracted by using rule-based reasoning function of ontology at the moment. As a practical example of architectural heritage, the Seokjojeon Hall is used to reveal a spatial experience scenario.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.6
• /
• pp.2012-2030
• /
• 2021

Named Entity Recognition (NER) is used to identify entity nouns in the corpus such as Location, Person and Organization, etc. NER is also an important basic of research in various natural language fields. The processing of Chinese NER has some unique difficulties, for example, there is no obvious segmentation boundary between each Chinese character in a Chinese sentence. The Chinese NER task is often combined with Chinese word segmentation, and so on. In response to these problems, we summarize the recognition methods of Chinese NER. In this review, we first introduce the sequence labeling system and evaluation metrics of NER. Then, we divide Chinese NER methods into rule-based methods, statistics-based machine learning methods and deep learning-based methods. Subsequently, we analyze in detail the model framework based on deep learning and the typical Chinese NER methods. Finally, we put forward the current challenges and future research directions of Chinese NER technology.

• Genomics & Informatics
• /
• v.17 no.4
• /
• pp.39.1-39.20
• /
• 2019

Analyzing patterns in data points embedded in linear and non-linear feature spaces is considered as one of the common research problems among different research areas, for example: data mining, machine learning, pattern recognition, and multivariate analysis. In this paper, data points are heterogeneous sets of biosequences (composite data points). A composite data point is a set of ordinary data points (e.g., set of feature vectors). We theoretically extend the derivation of the largest generalized eigenvalue-based distance metric D_ij(γ₁) in any linear and non-linear feature spaces. We prove that D_ij(γ₁) is a metric under any linear and non-linear feature transformation function. We show the sufficiency and efficiency of using the decision rule $\bar{{\delta}}_{{\Xi}i}$(i.e., mean of D_ij(γ₁)) in classification of heterogeneous sets of biosequences compared with the decision rules min_𝚵iand median_𝚵i. We analyze the impact of linear and non-linear transformation functions on classifying/clustering collections of heterogeneous sets of biosequences. The impact of the length of a sequence in a heterogeneous sequence-set generated by simulation on the classification and clustering results in linear and non-linear feature spaces is empirically shown in this paper. We propose a new concept: the limiting dispersion map of the existing clusters in heterogeneous sets of biosequences embedded in linear and nonlinear feature spaces, which is based on the limiting distribution of nucleotide compositions estimated from real data sets. Finally, the empirical conclusions and the scientific evidences are deduced from the experiments to support the theoretical side stated in this paper.

• Journal of the Korean School Mathematics Society
• /
• v.10 no.1
• /
• pp.91-111
• /
• 2007

This study investigated how high school students predict the rule, the sum of sequence for the concept of sequence, for the given patterns based on inductive approach using computers that provide dynamic functions and materials that are visual. Students for themselves were able to induce the formula without using the given formula in the textbook. Furthermore, this study examined how these technology and materials affect students' understanding of the concept of actual infinity for those who have the concept of the potential infinity which is the misconception of infinity in a infinity series. This study shows that students made a progress from the concept of potential infinity to that of actual infinity with technology and materials used I this study. Students also became interested in the use of computer and the visualized materials, further there was a change in their attitude toward mathematics.