• Journal of Intelligence and Information Systems
• /
• v.12 no.1
• /
• pp.107-123
• /
• 2006

We propose the Generalized Binary Second-order Recurrent Neural Networks(GBSRNNf) being equivalent to regular grammars and ?how the implementation of lexical analyzer recognizing the regular languages by using it. All the equivalent representations of regular grammars can be implemented in circuits by using GSBRNN, since it has binary-valued components and shows the structural relationship of a regular grammar. For a regular grammar with the number of symbols m, the number of terminals p, the number of nonterminals q, and the length of input string k, the size of the corresponding GBSRNN is $O(m(p+q)^2)$ and its parallel processing time is O(k) and its sequential processing time, $O(k(p+q)^2)$.

• Journal of Educational Research in Mathematics
• /
• v.16 no.4
• /
• pp.367-386
• /
• 2006

There are two strands for considering tile relationships between education and technology. One is the viewpoint of 'learning from computers' and the other is that of 'learning with computers'. In this paper, we call mathematics education with computers as 'computers and mathematics education' and this computer environments as microworlds. In this paper, we first suggest theoretical backgrounds ai constructionism, mathematization, and computer interaction. These theoretical backgrounds are related to students, school mathematics and computers, relatively As specific strategies to design a microworld, we consider a physical construction, fuctiionization, and internet interaction. Next we survey the different microworlds such as Logo and Dynamic Geometry System(DGS), and reform each microworlds for mathematical level-up of representation. First, we introduce the concept of action letters and its manipulation for representing turtle actions and recursive patterns in turtle microworld. Also we introduce another algebraic representation for representing DGS relation and consider educational moaning in dynamic geometry microworld. We design an integrating microworld between Logo and DGS. First, we design a same command system and we get together in a microworld. Second, these microworlds interact each other and collaborate to construct and manipulate new objects such as tiles and folding nets.

• Phonetics and Speech Sciences
• /
• v.12 no.4
• /
• pp.63-71
• /
• 2020

Speech recognition is one of the areas actively commercialized using deep learning and machine learning techniques. However, the majority of speech recognition systems on the market are developed on data with limited diversity of speakers and tend to perform well on typical adult speakers only. This is because most of the speech recognition models are generally learned using a speech database obtained from adult males and females. This tends to cause problems in recognizing the speech of the elderly, children and people with dialects well. To solve these problems, it may be necessary to retain big database or to collect a data for applying a speaker adaptation. However, this paper proposes that a new end-to-end speech recognition method consists of an acoustic augmented recurrent encoder and a transformer decoder with linguistic prediction. The proposed method can bring about the reliable performance of acoustic and language models in limited data conditions. The proposed method was evaluated to recognize Korean elderly and children speech with limited amount of training data and showed the better performance compared of a conventional method.

• Journal of Advanced Marine Engineering and Technology
• /
• v.41 no.1
• /
• pp.56-61
• /
• 2017

This paper presents a data prediction method and its application to haptic-based teleoperation systems. In general, time delays inevitably occur during data transmission in a network environment, which degrades the overall performance of haptic-based teleoperation systems. To address this situation, this paper proposes an autoregressive moving average (ARMA) model-based data prediction algorithm for estimating model parameters and predicting future data recursively in real time. The proposed method was applied to haptic data captured every 5 ms while bilateral haptic interaction was carried out by two users with an object in a virtual space. The results showed that the prediction performance of the proposed method had an error of less than 1 ms when predicting position-level data 100 ms ahead.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2021.06a
• /
• pp.116-116
• /
• 2021

자연 하천에서의 부유사 농도 계측은 주로 재래식 채집방식을 활용한 직접계측 방식에 의존하여 비용과 시간이 많이 소요되며 점 계측 방식으로 고해상도의 시공간 자료를 측정하기엔 한계가 존재한다. 이러한 한계점을 극복하기 위해 최근 위성영상과 드론을 활용하여 촬영된 다분광 혹은 초분광 영상을 통해 고해상도의 부유사 농도 시공간분포를 측정하는 기법에 대한 연구가 활발히 진행되고 있다. 하지만, 다른 하천 물리량 계측에 비해 부유사 계측 연구는 하천에 따라 부유사가 비균질적으로 분포하여 원격탐사를 통해 정확하고 전역적인 농도 분포를 재현하기는 어려운 실정이다. 이러한 부유사의 비균질성은 부유사의 입도분포, 광물특성, 침강성 등이 하천에서 다양하게 분포하기 때문이며 이로 인해 부유사는 지역별로 다양한 분광특성을 가지게 된다. 따라서, 본 연구에서는 이러한 영향을 고려한 전역적인 부유사 농도 예측 모형을 개발하기 위해 실내 실험을 통해 부유사 특성별 고유 분광 라이브러리를 구축하고 실규모 수로에서 다양한 부유사 조건에 대한 초분광 스펙트럼과 부유사 농도를 측정하는 실험을 수행하였다. 실제 부유사 농도는 광학 기반 센서인 LISST-200X와 샘플링을 통한 실험실 분석을 통해 계측되었으며, 초분광 스펙트럼 자료는 초분광 카메라를 통해 촬영한 영상에서 부유사 계측 지점에 대한 픽셀의 스펙트럼을 추출하여 구축하였다. 이렇게 생성된 자료들의 분광 다양성을 주성분 분석(Principle Component Analysis; PCA)를 통해 분석하였으며, 부유사의 입도 분포, 부유사 종류, 수온 등과의 상관관계를 통해 분광 특성과 가장 상관관계가 높은 물리적 인자를 규명하였다. 더불어 구축된 자료를 바탕으로 기계학습 기반 주요 특징 선택 알고리즘인 재귀적 특징 제거법 (Recursive Feature Elimination)과 기계학습기반 회귀 모형인 Support Vector Regression을 결합하여 초분광 영상 기반 부유사 농도 예측 모형을 개발하였으며, 이 결과를 원격탐사 계측 연구에서 일반적으로 사용되어 오던 최적 밴드비 분석 (Optimal Band Ratio Analysis; OBRA) 방법으로 도출된 회귀식과 비교하였다. 그 결과, 기존의 OBRA 기반 방법은 비선형성을 증가시켜도 좁은 영역의 파장대만을 고려하는 한계점으로 인해 부유사의 다양한 분광 특성을 반영하지 못하였으며, 본 연구에서 제시한 기계학습 기반 예측 모형은 420 nm~1000 nm에 걸쳐 폭 넓은 파장대를 고려함과 동시에 높은 정확도를 산출하였다. 최종적으로 개발된 모형을 적용해 다양한 유사 조건에 대한 부유사 시공간 분포를 매핑한 결과, 시공간적으로 고해상도의 부유사 농도 분포를 산출하는 것으로 밝혀졌다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2019.07a
• /
• pp.115-116
• /
• 2019

SVM (Support Vector Machine)은 견고성으로 인해 다양한 분류 문제에 적용 할 수 있는 효율적인 기계 학습 기술이다. 그러나 훈련 데이터의 수가 증가함에 따라 시간 복잡도가 급격히 증가하므로 대규모 데이터 세트의 경우 SVM이 비실용적이다. 본 논문에서는 SVM을 사용하여 중복 된 학습 데이터를 효율적으로 제거하는 새로운 병렬 평면(Parallel Hyperplane) 기법을 소개한다. 제안 기법에서 PH는 재귀 적으로 형성되는 반면 PH의 외부에 있는 데이터 포인트의 클러스터는 매 반복마다 제거된다. 시뮬레이션 결과 제안 기법은 기존의 클러스터링 기반 감축 기법과 SMO 기법에 비해 학습 시간을 크게 단축시키면서 데이터 축소 없이 분류의 정확성을 높일 수 있음을 확인 하였다.

• Annual Conference on Human and Language Technology
• /
• 2018.10a
• /
• pp.580-583
• /
• 2018

본 논문에서는 SeqGAN 모델을 사용하여 한국어 시를 자동 생성해 보았다. SeqGAN 모델은 문장 생성을 위해 재귀 신경망과 강화 학습 알고리즘의 하나인 정책 그라디언트(Policy Gradient)와 몬테카를로 검색(Monte Carlo Search, MC) 기법을 생성기에 적용하였다. 시 문장을 자동 생성하기 위한 학습 데이터로는 사랑을 주제로 작성된 시를 사용하였다. SeqGAN 모델을 사용하여 자동 생성된 시는 동일한 구절이 여러번 반복되는 문제를 보였지만 한국어 텍스트 생성에 있어 SeqGAN 모델이 적용 가능함을 확인하였다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.24 no.3
• /
• pp.349-354
• /
• 2020

The preliminary investigation to expand the green space requires a lot of cost and time. In this paper, we solve the problem by measuring the ratio of green space in a specific region through a convolutional neural network based the green space classification using Google Earth images. First, the proposed method collects various region images in Google Earth and learns them by using the convolutional neural network. The proposed method divides the image recursively to measure the green space ratio of the specific region, and it determines whether the divided image is green space using a trained convolutional neural network model, and then the green space ratio is calculated using the regions determined as the green space. Experimental results show that the proposed method shows high performance in measuring green space ratios in various regions.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.38 no.4
• /
• pp.363-373
• /
• 2020

During last decades numerous studies generating orthoimage have been carried out. Traditional methods require exterior orientation parameters of aerial images and precise 3D object modeling data and DTM (Digital Terrain Model) to detect and recover occlusion areas. Furthermore, it is challenging task to automate the complicated process. In this paper, we proposed a new concept of true orthoimage generation using DL (Deep Learning). DL is rapidly used in wide range of fields. In particular, GAN (Generative Adversarial Network) is one of the DL models for various tasks in imaging processing and computer vision. The generator tries to produce results similar to the real images, while discriminator judges fake and real images until the results are satisfied. Such mutually adversarial mechanism improves quality of the results. Experiments were performed using GAN-based Pix2Pix model by utilizing IR (Infrared) orthoimages, intensity from LiDAR data provided by the German Society for Photogrammetry, Remote Sensing and Geoinformation (DGPF) through the ISPRS (International Society for Photogrammetry and Remote Sensing). Two approaches were implemented: (1) One-step training with intensity data and high resolution orthoimages, (2) Recursive training with intensity data and color-coded low resolution intensity images for progressive enhancement of the results. Two methods provided similar quality based on FID (Fréchet Inception Distance) measures. However, if quality of the input data is close to the target image, better results could be obtained by increasing epoch. This paper is an early experimental study for feasibility of DL-based true orthoimage generation and further improvement would be necessary.

• The Transactions of the Korea Information Processing Society
• /
• v.13 no.4
• /
• pp.199-207
• /
• 2024

Local reservoirs are crucial sources for agricultural water supply, necessitating stable water level management to prepare for extreme climate conditions such as droughts. Water level prediction is significantly influenced by local climate characteristics, such as localized rainfall, as well as seasonal factors including cropping times, making it essential to understand the correlation between input and output data as much as selecting an appropriate prediction model. In this study, extensive multivariate data from over 400 reservoirs in Jeollabuk-do from 1991 to 2022 was utilized to train and validate a water level prediction model that comprehensively reflects the complex hydrological and climatological environmental factors of each reservoir, and to analyze the impact of each input feature on the prediction performance of water levels. Instead of focusing on improvements in water level performance through neural network structures, the study adopts a basic Feedforward Neural Network composed of fully connected layers, batch normalization, dropout, and activation functions, focusing on the correlation between multivariate input data and prediction performance. Additionally, most existing studies only present short-term prediction performance on a daily basis, which is not suitable for practical environments that require medium to long-term predictions, such as 10 days or a month. Therefore, this study measured the water level prediction performance up to one month ahead through a recursive method that uses daily prediction values as the next input. The experiment identified performance changes according to the prediction period and analyzed the impact of each input feature on the overall performance based on an Ablation study.