• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.178-185
• /
• 2021

Sequence to sequence(S2S) 기반 음성인식 후처리기를 훈련하기 위한 학습 데이터 구축을 위해 (음성인식 결과(speech recognition sentence), 전사자(phonetic transcriptor)가 수정한 문장(Human post edit sentence))의 병렬 말뭉치가 필요하며 이를 위해 많은 노동력(human-labor)이 소요된다. BackTranScription (BTS)이란 기존 S2S기반 음성인식 후처리기의 한계점을 완화하기 위해 제안된 데이터 구축 방법론이며 Text-To-Speech(TTS)와 Speech-To-Text(STT) 기술을 결합하여 pseudo 병렬 말뭉치를 생성하는 기술을 의미한다. 해당 방법론은 전사자의 역할을 없애고 방대한 양의 학습 데이터를 자동으로 생성할 수 있기에 데이터 구축에 있어서 시간과 비용을 단축 할 수 있다. 본 논문은 BTS를 바탕으로 제주어 도메인에 특화된 음성인식 후처리기의 성능을 향상시키기 위하여 모델 수정(model modification)을 통해 성능을 향상시키는 모델 중심 접근(model-centric) 방법론과 모델 수정 없이 데이터의 양과 질을 고려하여 성능을 향상시키는 데이터 중심 접근(data-centric) 방법론에 대한 비교 분석을 진행하였다. 실험결과 모델 교정없이 데이터 중심 접근 방법론을 적용하는 것이 성능 향상에 더 도움이 됨을 알 수 있었으며 모델 중심 접근 방법론의 부정적 측면 (negative result)에 대해서 분석을 진행하였다.

• Journal of the Korea Convergence Society
• /
• v.12 no.11
• /
• pp.109-117
• /
• 2021

Recently, the use of speech-based interfaces is increasing as a means for human-computer interaction (HCI). Accordingly, interest in post-processors for correcting errors in speech recognition results is also increasing. However, a lot of human-labor is required for data construction. in order to manufacture a sequence to sequence (S2S) based speech recognition post-processor. To this end, to alleviate the limitations of the existing construction methodology, a new data construction method called Back TranScription (BTS) was proposed. BTS refers to a technology that combines TTS and STT technology to create a pseudo parallel corpus. This methodology eliminates the role of a phonetic transcriptor and can automatically generate vast amounts of training data, saving the cost. This paper verified through experiments that data should be constructed in consideration of text style and domain rather than constructing data without any criteria by extending the existing BTS research.

• Annual Conference on Human and Language Technology
• /
• 2022.10a
• /
• pp.465-469
• /
• 2022

자동 음성 인식 (Automatic Speech Recognition) 기술이 발달함에 따라 자동 음성 인식 시스템의 성능을 높이기 위한 방법 중 하나로 자동 후처리기 연구(automatic post-processor)가 진행되어 왔다. 후처리기를 훈련시키기 위해서는 오류 유형이 포함되어 있는 병렬 말뭉치가 필요하다. 이를 만드는 간단한 방법 중 하나는 정답 문장에 오류를 삽입하여 오류 문장을 생성하여 pseudo 병렬 말뭉치를 만드는 것이다. 하지만 이는 실제적인 오류가 아닐 가능성이 존재한다. 이를 완화시키기 위하여 Back TranScription (BTS)을 이용하여 후처리기 모델 훈련을 위한 병렬 말뭉치를 생성하는 방법론이 존재한다. 그러나 해당 방법론으로 생성 할 경우 노이즈가 적을 수 있다는 관점이 존재하다. 이에 본 연구에서는 BTS 방법론과 인위적으로 노이즈 강도를 추가한 방법론 간의 성능을 비교한다. 이를 통해 BTS의 정량적 성능이 가장 높은 것을 확인했을 뿐만 아니라 정성적 분석을 통해 BTS 방법론을 활용하였을 때 실제 음성 인식 상황에서 발생할 수 있는 실제적인 오류를 더 많이 포함하여 병렬 말뭉치를 생성할 수 있음을 보여준다.

• Journal of Ginseng Research
• /
• v.23 no.1 s.53
• /
• pp.44-49
• /
• 1999

Cu/Zn superoxide dismutase (SOD1) is a protective enzyme responsible for the dismutat ion of superoxide radicals within the cell by converting superoxide radicals to oxygen and hydrogen peroxide, which is in turn changed to oxygen and water by catalase. Previously, we reported that the panaxadiol (PD) and its ginsenoside $Rb_2$ induced the expression of SOD1 gene through AP2 binding site and its induction. Here, we examined the effect of subfractions of panaxadiol ginsenosides, which were extracted from different parts of ginseng root that possess various ratios of panaxadiol to panaxatriol, on the induction of SOD1 gene expression. To explore this possibility, the upstream regulatory region of SOD1 was linked to the chloramphenicol acetyl transferase (CAT) structural gene and introduced into human hepatoma HepG2 cells. We observed that the transcriptional activation of SOD1 was proportional to the contents ratio of panaxadiol ginsensides. Consistent with this results, the total extract portion prepared from the finely-hairy root, which contains the higher ratio of panaxadiol to panaxatriol about 2.6, increased the SODl transcription about 3 fold. This results suggest that the panaxadiol fraction could induce the SOD1 and total extract of the ginseng finely-hairy root would be a useful material as a functional food for the SOD1 inducer.

• Journal of Life Science
• /
• v.34 no.4
• /
• pp.271-278
• /
• 2024

Redox factor (Ref)-1, a ubiquitously expressed protein, acts as a modulator of redox-sensitive tran- scription factors and as an endonuclease in the repair pathway of damaged DNA. However, the function of Ref-1 in the differentiation of monocytes into macrophages has not been defined. In this study, we investigated the effects of Ref-1 on the monocyte differentiation process using the human monocytic cell line THP-1. The differentiation agent PMA increased cell adhesion over time and showed a sig- nificant increase in phagocytic function but decreased the intracellular amount of Ref-1. Ref-1 inhibitor E3330 and Ref-1 knockdown using the siRNA technique reduced cell adhesion and the expression of differentiation markers, such as CD14, ICAM-1, and CD11b, by PMA stimulation. This means that the role of Ref-1 is absolutely necessary in the initial process of differentiating THP-1 cells stimulated by PMA. Next, the distribution of Ref-1 was examined in the cytoplasm and nucleus of THP-1 cells stimulated with PMA. Surprisingly, PMA stimulation resulted in the rapid translocation of Ref-1 to the nucleus. To prove that movement of Ref-1 to the nucleus is required for monocyte differentiation, a Ref-1 vector with the nuclear localization sequence (NLS) deleted was used. As a result, overexpression of ∆NLS Ref-1, which restricted movement to the nucleus, suppressed the expression of differentiation markers and notably reduced phagocytic function in PMA-stimulated THP-1 cells. In conclusion, these data suggest that the differentiation of monocytic THP-1 cells requires Ref-1 nuclear translocation during the initial process of biochemical events following stimulation from PMA.