• Journal of the Korea Computer Graphics Society
• /
• v.27 no.5
• /
• pp.73-79
• /
• 2021

Recently, in the field of computer animation, a method for generating motion using deep learning has been studied away from conventional finite-state machines or graph-based methods. The expressiveness of the network required for learning motions is more influenced by the diversity of motion contained in it than by the simple length of motion to be learned. This study aims to find an efficient network structure when the types of motions to be learned are diverse. In this paper, we train and compare three types of networks: basic fully-connected structure, mixture of experts structure that uses multiple fully-connected layers in parallel, recurrent neural network which is widely used to deal with seq2seq, and transformer structure used for sequence-type data processing in the natural language processing field.

• Annual Conference of KIPS
• /
• 2011.04a
• /
• pp.1260-1263
• /
• 2011

최근 차세대 시퀀싱 (Next Generation Sequencing, NGS) 기술이 발전하면서 DNA, RNA 등의 시퀀싱 데이터를 이용한 유전체 분석 방식에 관한 연구가 활발히 이루어지고 있다. 차세대 시퀀싱 데이터를 이용한 유전체 분석 방식은 마이크로어레이 혹은 EST/cDNA 데이터를 이용한 기존의 분석 방식에 비하여 비용이 적게 들고 정확한 결과를 얻을 수 있다는 장점이 있다. 그러나 이 들 DNA, RNA 시퀀싱 데이터는 각 시퀀스의 길이가 짧고 전체 용량은 매우 커서 이 들 데이터로부터 정확한 분석 결과를 추출하는 데에 많은 어려움이 있다. 본 연구에서는 클라우드 컴퓨팅 기술을 기반으로 하여 대용량의 RNA 시퀀싱 데이터를 고속으로 처리하는 병렬 SNP 추출 알고리즘을 제안한다. 전체 게놈 데이터 중 유전자 영역만을 high coverage로 시퀀싱하여 얻어지는 RNA 시퀀싱 데이터는 유전자 변이 추출을 목적으로 분석되며, SNP(Single Nucleotide Polymorphism)와 같은 유전자 변이는 질병의 원인 규명 및 치료법 개발에 직접 이용된다. 제안된 알고리즘은 동시에 실행되는 다수의 Map/Reduce 함수에 의해서 대규모 RNA 시퀀스를 병렬로 처리하며, 레퍼런스 시퀀스에 매핑된 각 염기의 출현 빈도와 품질점수를 이용하여 SNP를 추출한다. 또한 이 들 SNP 추출 결과에 대한 시각적 분석 도구를 제공하여 SNP 추출 과정 및 근거를 시각적으로 확인/검증할 수 있도록 지원한다.

• 한국해양학회지
• /
• v.27 no.1
• /
• pp.19-34
• /
• 1992

Marine seismic refraction surveys have been carried out by Korea Institute of Geology, Mining and Materials(KIGAM) since 1984. The recording of refraction data was based on analog instrumentation. Therefore the resolution of refraction data was not good enough to distinguish many layers. The objective of the interpretation of seismic refraction data is the determination of intervals and critically refracted seismic wave propagation velocities through the layers beneath the sea floor. To determine intervals and velocities precisely, the resolution of refraction data should be enhanced. The intent of the study is to improve the quality of shallow marine refraction data by the digital technique using microcomputer- based acquisition and processing system. The system consists of an IBM AT microcomputer clone, an analog-digital(A/D) converter. A mass storage unit and a parallel processing board. The A/D converter has 12 bits of precision and 250 kHz of conversion rate. The magneto-optical disk drive is used for the mass storage of seismic refraction data. Shallow marine seismic refraction surveys have been carried out using the system at 6 locations off Ulsan and Pusan area. The refraction data were acquired by the radio sonobuoy. The refraction profiles have been produced by the laser printer with 300 dpi resolution after the basic computer processing. 5-9 layers were interpreted from digital refraction profiles, whereas 2-4 layers were interpreted from analog refraction profiles. the propagation velocities of sediments were interpreted as 1.6-2.1 km/sec. The propagation velocities of acoustic basement were interpreted as 2.4-2.7 km/sec off Ulsan area, 4.8 km/sec off Pusan area.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.49 no.5
• /
• pp.15-20
• /
• 2012

In this paper, we show new method to find the error locations of received Reed-Solomon code word. New design is much faster and has much simpler logic circuit than the former design method. This optimization was possible by very simplified square/$X^4$ calculating circuit, parallel processing and not using the very complex Divider. The Reed Solomon decoder using this new Chien Machine can be applicated for data protection of almost all digital communication and consumer electronic devices.

• Atmosphere
• /
• v.23 no.2
• /
• pp.231-235
• /
• 2013

We parallelized WRF major physics routines for Nvidia GP-GPUs with CUDA Fortran. GP-GPUs are originally designed for graphic processing, but show high performance with low electricity for calculating numerical models. In the CUDA environment, a data domain is allocated into thread blocks and threads in each thread block are computing in parallel. We parallelized the WRF program to use of thread blocks efficiently. We validated the GP-GPU program with the original CPU program, and the WRF model using GP-GPUs shows efficient speedup.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.11C
• /
• pp.649-654
• /
• 2011

In this paper, we show new method to find number of errors in the Reed-Solomon decoder. New design is much faster and has much simpler logic circuit than the former design method. This optimization was possible by very simplified square calculating circuit and parallel processing. The microcontroller of this Reed Solomon decoder can be used for data protection of almost all digital communication and consumer electronic devices.

• Journal of the Korean Institute of Telematics and Electronics A
• /
• v.29A no.11
• /
• pp.70-77
• /
• 1992

Generally, the conventional tracking algorithms are very limited in the practical applications because of that the computation load is exponentially increased as the number of targets being tracked is increase. Recently, to overcome this kind of limitation, some new tracking methods based on neural network algorithms which have learning and parallel processing capabilities are introduced. By application of neural networks to multi-target tracking problems, the tracking system can be made computationally independent of the number of objects being tracked, through their characteristics of massive parallelism and dense interconnectivity. In this paper, a new neural network tracking algorithm, which has capability of adaptive target tracking with little increase of the amount of calculation under the clutter and noisy environments, is suggested and the possibility of real-time multi-target tracking system based on neural networks is also demonstrated through some good computer simulation results.

• Proceedings of the KIEE Conference
• /
• 1991.11a
• /
• pp.39-42
• /
• 1991

The Application of Artificial Neural Network(ANN) to forecast a load in a power system is investigated. The load forecasting is important in the electric utility industry. This technique, methodology based on the fact that parallel structure can process very fast much information is a promising approach to a load forecasting. ANN that is highly interconnected processing element in a hierachy activated by the each input. The load pattern can be divided distinctively into two patterns, that is, weekday and weekend. ANN is composed of a input layer, several hidden layers, and a output layer and the past data is used to activate input layer. The output of ANN is the load forecast for a given day. The result of this simulation can be used as a reference to a electric utility operation.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.8 no.1
• /
• pp.127-137
• /
• 2002

This paper deals with developing a Web-based Solver NRO(Network Reliability Optimizer) for solving three classes of reliability redundancy optimization problems which are generated in series systems. parallel systems and complex systems. Inputs of NRO consisted in four parts. that is, user authentication. system selection. input data and confirmation. After processing of inputs through internet, NRO provides conveniently the optimal solutions for the given problems on the Web-site. To alleviate the risks of being trapped in a local optimum, HH(Hybrid-Heuristic) algorithm is incorporated in NRO for solving the given three classes of problems, and moderately combined GA(Genetic Algorithm) with the modified SA(Simulated Annealing) algorithm.

• Journal of the Korea Computer Industry Society
• /
• v.4 no.4
• /
• pp.501-508
• /
• 2003

In this paper A new theory of linear loop transformation called Elimination of Data Dependency(BDD) is presented. The current framework of linear loop transformation cannot identify a significant fraction of parallelism. For this reason, a method to extract the maximum loop parallelism in perfect nested loops is presented. This technique is applicable to general loop nests where the dependence include both distance and directions.