• THE INTERNATIONAL COMMERCE & LAW REVIEW
• /
• v.66
• /
• pp.219-240
• /
• 2015

Developing countries have faced serious difficulties in implementing customs procedures. For example, they have experienced heavy administrative constraints such as lack of qualified personnel and non-computerization. Some importers in developing countries have tried to undervalue or overvalue the imported goods and misclassify them with the target of lowering the incidence of customs duties. In order to address these problems many of developing countries often require the exporter to get a Clean Report of Findings(CRF) issued by a preshipment inspection entity. This study tries to reply to the following questions with which the exporters are faced in the process of issuance of CRF. First, what are the provisions and regulations that the exporters have to note in the event a CRF is issued by the inspection entity? Second, how do the exporters safeguard their own rights in case of price verification? Third, how to protect the confidential business information to be submitted by the exporters in order to get a CRF? Fourth, how can the exporters respond to the improper conducts that the inspection entities have done?

• Annual Conference on Human and Language Technology
• /
• 2016.10a
• /
• pp.3-8
• /
• 2016

형태소 분석기는 많은 자연어 처리 영역에서 필수적인 언어 도구로 활용되기 때문에 형태소에 대한 품사를 결정하는 것은 매우 중요하다. 최근 음절 기반으로 형태소의 품사를 태깅하는 방법에 대한 연구들이 많이 진행되고 있다. 음절 단위 형태소 분석은 음절 단위로 분리된 형태소에 대해서 기계학습을 이용하여 분리된 음절 단위로 품사를 태깅하는 단계를 가진다. 본 논문에서는 기존의 CRF를 이용한 음절 단위 품사 태깅 방법을 개선하기 위해 bi-LSTM-CRFs를 이용한 방법을 제안한다. 또한, bi-LSTM-CRFs의 입력을 음절의 품사 분포 벡터를 이용해 확장함으로써 음절 단위 품사 태깅의 성능을 향상 시켰다.

• Annual Conference on Human and Language Technology
• /
• 2015.10a
• /
• pp.85-89
• /
• 2015

본 논문에서는 시퀀스 레이블링 문제에 적합하다고 알려진 Long Short Term Memory Recurrent Neural Network에 아웃풋간의 의존관계를 추가한 LSTM-CRF(Conditional Random Field)를 이용하여 생명과학분야 개체명 인식 시스템을 구축하였다. 학습 및 평가를 위해 BioNLP 2011-st REL data를 개체명 인식 실험에 사용하였으며, 실험결과 LSTM-CRF를 사용한 시스템은 81.83의 F1-score를 기록해, 기존의 시스템인 "BANNER"의 F1-score 81.96과 비슷한 성능을 보였다.

• Annual Conference on Human and Language Technology
• /
• 2015.10a
• /
• pp.194-197
• /
• 2015

Long Short-term Memory Network(LSTM) 기반 Recurrent Neural Network(RNN)는 순차 데이터를 모델링 할 수 있는 딥 러닝 모델이다. 기존 RNN의 그래디언트 소멸 문제(vanishing gradient problem)를 해결한 LSTM RNN은 멀리 떨어져 있는 이전의 입력 정보를 볼 수 있다는 장점이 있어 음성 인식 및 필기체 인식 등의 분야에서 좋은 성능을 보이고 있다. 또한 LSTM RNN 모델에 의존성(전이 확률)을 추가한 LSTM CRF모델이 자연어처리의 한 분야인 개체명 인식에서 우수한 성능을 보이고 있다. 본 논문에서는 한국어 문장의 지배소가 문장 후위에 나타나는 점에 착안하여 Backward 방식의 LSTM CRF 모델을 제안하고 이를 한국어 의미역 결정에 적용하여 기존 연구보다 더 높은 성능을 얻을 수 있음을 보인다.

• Journal of Integrative Natural Science
• /
• v.11 no.1
• /
• pp.19-24
• /
• 2018

CRFR is involved in the pathophysiology of various disorders including depression, stress, anxiety, post-traumatic stress disorder, and addiction. The discovery of novel and structurally diverse CRF1 receptor inhibitors becomes essential. In this study, we have performed molecular docking of CRF1R with the derivatives of 8-substituted-2-aryl-5-alkylaminoquinolines as CRF1R inhibitors. The antagonist molecules were optimized and docked into the binding site of the receptor. On analysing the docked complexes we have identified that the residues HIS214, THR215, ARG227, ARG1008, LYS1060 and ASP1061 are important in forming hydrogen bond with the inhibitors. Further studies on these residues could reveal important structural features required for the formation of CRF1R-inhibitor complex and thus in the discovery of novel and potent inhibitors.

• Journal of Advanced Information Technology and Convergence
• /
• v.10 no.2
• /
• pp.139-151
• /
• 2020

Entity name recognition is a part of information extraction that extracts entity names from documents and classifies the types of extracted entity names. Entity name recognition technologies are widely used in natural language processing, such as information retrieval, machine translation, and query response systems. Various deep learning-based models exist to improve entity name recognition performance, but studies that compared and analyzed these models on Korean data are insufficient. In this paper, we compare and analyze the performance of CRF, LSTM-CRF, BiLSTM-CRF, and BERT, which are actively used to identify entity names using Korean data. Also, we compare and evaluate whether embedding models, which are variously used in recent natural language processing tasks, can affect the entity name recognition model's performance improvement. As a result of experiments on patent data and Korean corpus, it was confirmed that the BiLSTM-CRF using FastText method showed the highest performance.

• Annual Conference on Human and Language Technology
• /
• 2019.10a
• /
• pp.34-36
• /
• 2019

기존 딥 러닝을 이용한 형태소 분석 및 품사 태깅(Part-Of-Speech tagging)은 feed-forward neural network에 CRF를 결합하는 방법이나 sequence-to-sequence 모델을 이용한 방법 등의 다양한 모델들이 연구되었다. 본 논문에서는 한국어 형태소 분석 및 품사 태깅을 수행하기 위하여 최근 자연어처리 태스크에서 많은 성능 향상을 보이고 있는 BERT를 기반으로 한 음절 단위 LSTM-CRF 모델을 제안한다. BERT는 양방향성을 가진 트랜스포머(transformer) 인코더를 기반으로 언어 모델을 사전 학습한 것이며, 본 논문에서는 한국어 대용량 코퍼스를 어절 단위로 사전 학습한 KorBERT를 사용한다. 실험 결과, 본 논문에서 제안한 모델이 기존 한국어 형태소 분석 및 품사 태깅 연구들 보다 좋은 (세종 코퍼스) F1 98.74%의 성능을 보였다.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.571-575
• /
• 2021

감성분석이란 텍스트에 들어있는 의견이나 감성, 평가, 태도 등의 주관적인 정보를 컴퓨터를 통해 분석하는 과정이다. 본 논문은 다양한 감성분석 실험 중 감성이 드러나는 부분을 파악하여 서술어 중심의 구 혹은 절 단위로 감성 표현 영역을 추출하는 모델을 개발하고자 한다. 제안하는 모델은 BERT에 classification layer와 CRF layer를 결합한 것이고 baseline은 일반 BERT 모델이다. 실험 결과는 기존의 baseline 모델의 f1-score이 33.44%이고 제안한 BERT+CRF 모델의 f1-score이 40.99%이다. BERT+CRF 모델이 7.55% 더 좋은 성능을 보인다.

• The Korean Journal of Pain
• /
• v.33 no.2
• /
• pp.183-191
• /
• 2020

Background: Opioids can present intolerable adverse side-effects to patients who use these analgesics to mitigate chronic pain. In this retrospective analysis, cooled radiofrequency (CRF) denervation was evaluated to provide pain and disability relief and reduce opioid use in patients with sacroiliac joint (SIJ) derived low back pain (LBP). Methods: Twenty-seven patients with pain from SIJ refractory to conservative treatments, and taking opioids chronically (> 3 mo), were included. Numeric rating scale (NRS) and Oswestry disability index (ODI) scores were collected at 1, 6, and 12 months post-procedure. Opioid use between baseline and each follow-up visit was compared for the entire group and for those who experienced successful (pain reduction ≥ 50% of baseline value) or unsuccessful CRF denervation. Results: Severe initial mean pain (NRS score: 7.7 ± 1.0) and disability (ODI score: 50.1 ± 9.0), and median opioid use (morphine equivalent daily dose: 40 ± 37 mg) were significantly reduced up to 12 months post-intervention. CRF denervation was successful in 44.4% of the patients at 12 months. Regardless of procedure success, patients demonstrated similar opioid reductions and changes in opioid use at 12 months. Two patients (7.4%) experienced neuritis following CRF denervation. Conclusions: CRF denervation of the SIJ can safely elicit pain and disability relief, and reduce opioid use, regardless of intervention success. Future studies may support CRF denervation as a dependable therapy to alleviate opioid use in patients with SIJ-derived LBP and show that opioid use measurements can be a surrogate indicator of pain.

• Herbal Formula Science
• /
• v.18 no.1
• /
• pp.155-167
• /
• 2010

Objectives : In order to investigate the antidepressant effects of Forsythiae Fructus(FF), we performed the Forced Swimming Test(FST). Also the expressions of corticotropin-releasing factor (CRF), c-Fos and tyrosine hydroxylase(TH) were measured by immunohistochemical method at paraventricular nucleus(PVN), ventral tegmental area(VTA) and locus coeruleus(LC). Methods : Spraque-Dawley rats were administered FF extract(100 mg/kg, 400 mg/kg) intragastrically three times prior to the FST. Results : The duration of immobility in the FST was significantly decreased in the FF 100 mg/kg, 400 mg/kg groups(p<0.05). The expression of CRF was significantly reduced in the FF 100 mg/kg and 400 mg/kg groups(p<0.001). c-Fos expression was significantly decreased at PVN in the FF 100 mg/kg group(p=0.069). TH expression at VTA was significantly increased in the FF 100 mg/kg and 400 mg/kg groups(p<0.05). TH expression at LC was not significantly changed(p=0.346). Conclusion : According to the results, it can be suggested that Forsythiae Fructus has antidepressant effect via the decreased immobility through the reduction of CRF and c-Fos expression at PVN.