• 한국정보교육학회:학술대회논문집
• /
• 2021.08a
• /
• pp.425-431
• /
• 2021

In this study, in order to develop a manual for the AI convergence education competency strengthening project, project personnel were interviewed and an overall analysis of the business process was made. As a result of analysis, the manual includes methods to solve problems currently being raised in the field such as recruitment of graduate students, education program management, academic registration management, and budget management. In order to increase the continuity and efficiency of the project, it is necessary to build an administrative management site that provides information on the overall project, to strengthen publicity for general teachers about AI convergence education, and meet the educational needs and level of enrolled students.

• The Journal of Korea Robotics Society
• /
• v.18 no.1
• /
• pp.122-126
• /
• 2023

An unexpected occlusion in a real life, not in a laboratory, can be more fatal to neural networks than expected. In addition, it is virtually impossible to create a network that learns all the environmental changes as well as occlusions. Therefore, we propose an alternative approach in which the architecture and number of parameters remain unchanged while adapting to occlusion circumstances. Learning method with the term Conversion Learning classifies them more robustly by converting the weights from various occlusion situations. The experiments on MNIST dataset showed a 3.07 [%p] performance improvement over the baseline CNN model in a situation where most objects are occluded and unknowing what occlusion will appear in advance. The experimental results suggest that Conversion Learning is an efficient method to respond to environmental changes such as occluded images.

• Journal of Ginseng Research
• /
• v.40 no.4
• /
• pp.366-374
• /
• 2016

Background: In this study, we screened and identified an endophyte JG09 having strong biocatalytic activity for ginsenosides from Platycodon grandiflorum, converted ginseng total saponins and ginsenoside monomers, determined the source of minor ginsenosides and the transformation pathways, and calculated the maximum production of minor ginsenosides for the conversion of ginsenoside Rb1 to assess the transformation activity of endophyte JG09. Methods: The transformation of ginseng total saponins and ginsenoside monomers Rb1, Rb2, Rc, Rd, Rg1 into minor ginsenosides F2, C-K and Rh1 using endophyte JG09 isolated by an organizational separation method and Esculin-R2A agar assay, as well as the identification of transformed products via TLC and HPLC, were evaluated. Endophyte JG09 was identified through DNA sequencing and phylogenetic analysis. Results: A total of 32 ${\beta}$-glucosidase-producing endophytes were screened out among the isolated 69 endophytes from P. grandiflorum. An endophyte bacteria JG09 identified as Luteibacter sp. effectively converted protopanaxadiol-type ginsenosides Rb1, Rb2, Rc, Rd into minor ginsenosides F2 and C-K, and converted protopanaxatriol-type ginsenoside Rg1 into minor ginsenoside Rh1. The transformation pathways of major ginsenosides by endophyte JG09 were as follows: $Rb1{\rightarrow}Rd{\rightarrow}F2{\rightarrow}C-K$; $Rb2{\rightarrow}C-O{\rightarrow}C-Y{\rightarrow}C-K$; $Rc{\rightarrow}C-Mc1{\rightarrow}C-Mc{\rightarrow}C-K$; $Rg1{\rightarrow}Rh1$. The maximum production rate of ginsenosides F2 and C-K reached 94.53% and 66.34%, respectively. Conclusion: This is the first report about conversion of major ginsenosides into minor ginsenosides by fermentation with P. grandiflorum endophytes. The results of the study indicate endophyte JG09 would be a potential microbial source for obtaining minor ginsenosides.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.7
• /
• pp.303-314
• /
• 2023

Voice conversion, a technology that allows an individual's speech data to be regenerated with the acoustic properties(tone, cadence, gender) of another, has countless applications in education, communication, and entertainment. This paper proposes an approach based on the StarGAN-VC model that generates realistic-sounding speech without requiring parallel utterances. To overcome the constraints of the existing StarGAN-VC model that utilizes one-hot vectors of original and target speaker information, this paper extracts feature vectors of target speakers using a pre-trained version of Rawnet3. This results in a latent space where voice conversion can be performed without direct speaker-to-speaker mappings, enabling an any-to-any structure. In addition to the loss terms used in the original StarGAN-VC model, Wasserstein distance is used as a loss term to ensure that generated voice segments match the acoustic properties of the target voice. Two Time-Scale Update Rule (TTUR) is also used to facilitate stable training. Experimental results show that the proposed method outperforms previous methods, including the StarGAN-VC network on which it was based.