http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
- 中文 을 입력하시려면 zhongwen을 입력하시고 space를누르시면됩니다.
- 北京 을 입력하시려면 beijing을 입력하시고 space를 누르시면 됩니다.
RISS 인기검색어
- 검색키워드
- 저자 : Hideki Fujinami (검색결과 2 건)
- 무료
- 기관 내 무료
- 유료
-
Deep Learning with Data Augmentation to Add Data Around Classification Boundaries
Hideki Fujinami,Gendo Kumoi,Masayuki Goto 대한산업공학회 2021 Industrial Engineeering & Management Systems Vol.20 No.3
Data augmentation methods are used as a technique to improve generalization by increasing the number of training data in image classification. However, most of these methods are not a data driven algorithm, the degree of improvement of generalization ability by performing these data augmentation methods differs between the domains of image data for training. Generative models are researched to use for augmenting data recently. In particular, Generative Adversarial Networks (GANs) (Goodfellow et al., 2014) that can generate clean image get attention as an excellent innovation in machine learning. As GANs extension method, there is a method called CGANs (Mirza and Osindero, 2014) that can be used for data augmentation. When enough training data for each class are not prepared for classification model, the same is true for training CGANs. In such case, CGAN generates noisy images. This makes a classification model to underfit to the original training data. Moreover, when a CGAN approximates the training data distribution, the CGAN generates new training data in the same region where training data densely exist. In such case, augmented data can’t reduce overfitting on the original training data. Therefore, our research contributes to augment data which meets these two requirements. In this study, we propose a method to generate data by the class specific GAN with small training data and selectively add generated data to the training data set that improves classification accuracy by using the entropy of the classification model. The feature of the proposed method is that it focuses on the positional relationship between data and the classification hyperplane in deep learning. In the proposed method, the entropy of the classification model is used to measure the positional relationship between the classification boundary and the data. As a result, the generalization performance is improved by adding the data around the classification boundary as new training data.
-
Matsumoto, Takahiro,Ohkubo, Kei,Honda, Kaoru,Yazawa, Akiko,Furutachi, Hideki,Fujinami, Shuhei,Fukuzumi, Shunichi,Suzuki, Masatatsu American Chemical Society 2009 JOURNAL OF THE AMERICAN CHEMICAL SOCIETY - Vol.131 No.26
<P>A (mu-eta(2):eta(2)-peroxo)dicopper(II) complex, [Cu(2)(H-L)(O(2))](2+) (1-O(2)), supported by the dinucleating ligand 1,3-bis[bis(6-methyl-2-pyridylmethyl)aminomethyl]benzene (H-L) is capable of initiating C-H bond activation of a variety of external aliphatic substrates (SH(n)): 10-methyl-9,10-dihydroacridine (AcrH(2)), 1,4-cyclohexadiene (1,4-CHD), 9,10-dihydroanthracene (9,10-DHA), fluorene, tetralin, toluene, and tetrahydrofuran (THF), which have C-H bond dissociation energies (BDEs) ranging from approximately 75 kcal mol(-1) for 1,4-CHD to approximately 92 kcal mol(-1) for THF. Oxidation of SH(n) afforded a variety of oxidation products, such as dehydrogenation products (SH((n-2))), hydroxylated and further-oxidized products (SH((n-1))OH and SH((n-2))=O), dimers formed by coupling between substrates (H((n-1))S-SH((n-1))) and between substrate and H-L (H-L-SH((n-1))). Kinetic studies of the oxidation of the substrates initiated by 1-O(2) in acetone at -70 degrees C revealed that there is a linear correlation between the logarithms of the rate constants for oxidation of the C-H bonds of the substrates and their BDEs, except for THF. The combination of this correlation and the relatively large deuterium kinetic isotope effects (KIEs), k(2)(H)/k(2)(D) (13 for 9,10-DHA, approximately > 29 for toluene, and approximately 34 for THF at -70 degrees C and approximately 9 for AcrH(2) at -94 degrees C) indicates that H-atom transfer (HAT) from SH(n) (SD(n)) is the rate-determining step. Kinetic studies of the oxidation of SH(n) by cumylperoxyl radical showed a correlation similar to that observed for 1-O(2), indicating that the reactivity of 1-O(2) is similar to that of cumylperoxyl radical. Thus, 1-O(2) is capable of initiating a wide range of oxidation reactions, including oxidation of aliphatic C-H bonds having BDEs from approximately 75 to approximately 92 kcal mol(-1), hydroxylation of the m-xylyl linker of H-L, and epoxidation of styrene (Matsumoto, T.; et al. J. Am. Chem. Soc. 2006, 128, 3874).</P>
연관 검색어 추천
이 검색어로 많이 본 자료
활용도 높은 자료
