http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
김원욱(Kim Weon Ook),조용현(Cho Yong Hyun),김영일(Kim Young Il),강인구(Kang In Ku) 한국정보처리학회 1997 정보처리학회논문지 Vol.4 No.11
This paper proposes an efficient method for improving the training performance of the neural networks using a hybrid of conjugate gradient backpropagation algorithm and dynamic tunneling backpropagation algorithm. The conjugate gradient backpropagation algorithm, which is the fast gradient algorithm, is applied for high speed optimization. The dynamic tunneling backpropagation algorithm, which is the deteministic method with tunneling phenomenon, is applied for global optimization. Converging to the local minima by using the conjugate gradient backpropagation algorithm, the new initial point for escaping the local minima is estimated by dynamic tunneling backpropagation algorithm. The proposed method has been applied to the parity check and to pattern classfication. The simulation results show that the performance of proposed method is superior to those of gradient descent backpropagation algorithm and a hybrid of gradient descent and dynamic thunneling backpropagation algorithm, and the new algorithm converges more often to the global minima than gradient descent backpropagation algorithm.
Kim, Suk-Weon,Ban, Sung-Hee,Jeong, Soon-Chun,Chung, Hwa-Jee,Ko, Suk-Min,Yoo, Ook-Joon,Liu, Jang R. Korean Society for Biotechnology and Bioengineerin 2007 Biotechnology and Bioprocess Engineering Vol.12 No.6
When whole cell extracts are subjected to proton nuclear magnetic resonance spectroscopy $(^1H\;NMR)$, metabolite profiles are generated that contain overlapping signals of the majority of compounds within the extract. In order to determine whether pattern recognition based on the metabolite profiles of higher plants is able to genetically discriminate between plants, we analyzed leaf samples of eight cultivars of Catharanthus roseus by $^1H$ NMR. Hierarchical dendrograms, based on the principal component analysis of the $^1H$ NMR total, aliphatic, carbohydrate, and aromatic region data, revealed possible relationships between the cultivars. The dendrogram based on the aromatic region data was in general agreement with the genetic relationships determined by conventional DNA fingerprinting methods. Secologanin and polyphenols were assigned to the signals of the $^1H$ NMR spectra, and contributed most profoundly to the discrimination between cultivars. The overall results indicate that the genetic relationships between C. roseus cultivars are reflected in the differences of the aromatic compounds in the leaves.
Kim Suk-Weon,Ban Sung-Hee,Yoo Ook-Joon,Liu Jang-Ryol The Korean Society for Biotechnology and Bioengine 2006 Biotechnology and Bioprocess Engineering Vol.11 No.1
When whole cells are subjected to pyrolysis gas chromatography/mass spectrometry (Py-GC/MS) analysis, it provides biochemical profiles containing overlapping signals of the majority of compounds. To determine marker compounds that discriminate embryogenic calluses from nonembryogenic calluses, samples of embryogenic and nonembryogenic calluses of five higher plant species were subjected to Py-GC/MS. Genetic programming of Py-GC/MS data was able to discriminate embryogenic calluses from nonembryogenic calluses. The content ratio of 5-meyhyl-2-furancarboxaldehyde and 5-(hydroxymethyl)-2-furancarboxaldehyde was greater in nonembryogenic calluses than in embryogenic calluses. However, the content ratio of phenol, p-cresol, and $^1H-indole$ in embryogenic calluses was 1.2 to 2.4 times greater than the ratio in nonembryogenic calluses. These pyrolysates seem to be derived from the components of the cell walls, which suggests that differences in cell wall components or changes in the architecture of the cell wall playa crucial role in determining the embryogenic competence of calluses.
김원욱,조용현,김영일,강인구,Kim, Weon-Ook,Cho, Yong-Hyun,Kim, Young-Il,Kang, In-Ku 한국정보처리학회 1997 정보처리논문지 Vol.4 No.11
본 논문에서는 공액기울기법과 터널링 시스템을 조합사용하여 신경망의 학습성능을 향상시킬 수 있는 효율적인 방법을 제안하였다. 빠른 수렴속도의 학습을 위하여 공액 기울기법에 기초한 후향전파 알고리즘을 사용하였고, 국소최적해를 만났을 때 이를 벗어난 다른 연결가중치의 설정을 위해 동적터널링 시스템에 기초한 후향전파 알고리즘을 조합한 학습 알고리즘을 적용하였다. 제안된 방법을 패리티 검사 및 패턴분류 문제에 각각 적용하여 기존의 기울기 하강법에 기초한 후향전파 알고리즘 및 기울기 하강법과 동적터널링 시스템을 조합한 후향전파 알고리즘방법의 결과와 비교 고찰하여 제안된 방법이 다른 방법들 보다 학습성능에서 우수함을 나타내었다. This Paper Proposes an efficient method for improving the training performance of the neural networks using a hybrid of conjugate gradient backpropagation algorithm and dynamic tunneling backpropagation algorithm The conjugate gradient backpropagation algorithm, which is the fast gradient algorithm, is applied for high speed optimization. The dynamic tunneling backpropagation algorithm, which is the deterministic method with tunneling phenomenon, is applied for global optimization. Conversing to the local minima by using the conjugate gradient backpropagation algorithm, the new initial point for escaping the local minima is estimated by dynamic tunneling backpropagation algorithm. The proposed method has been applied to the parity check and the pattern classification. The simulation results show that the performance of proposed method is superior to those of gradient descent backpropagtion algorithm and a hybrid of gradient descent and dynamic tunneling backpropagation algorithm, and the new algorithm converges more often to the global minima than gradient descent backpropagation algorithm.