• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.326-327
• /
• 2021

Newly, big data utilization has been widely interested in a wide variety of industrial fields. Big data analysis is the process of discovering meaningful new correlations, patterns, and trends in large volumes of data stored in data stores and creating new value. Thus, most big data analysis technology methods include data mining, machine learning, natural language processing, and pattern recognition used in existing statistical computer science. Also, using the R language, a big data tool, we can express analysis results through various visualization functions using pre-processing text data. The data used in this study were analyzed for 29 papers in a specific journal. In the final analysis results, the most frequently mentioned keyword was "Research", which ranked first 743 times. Therefore, based on the results of the analysis, the limitations of the study and theoretical implications are suggested.

• Proceedings of the Korean Society for Language and Information Conference
• /
• 2002.02a
• /
• pp.320-329
• /
• 2002

The issues in polysemy with respect to the verbs in WordNet will be discussed in this paper. The hypernymy/hyponymy structure of the multiple senses is observed when we try to build a bilingual network for Chinese and English. There are several types of polysemic patterns and a co-hypernym may have the same word form as its subordinates. Fellbaum (2000) dubbed autotroponymy that the verbs linked by mailer relation share the same verb form. However, her syntactic criteria seem not compatible to the hierarchies in WN. Either the criteria or the network should be reconducted. For most verbs in WN 1.7, polysemous relations are unlikely to extend over 3 levels of IS-A relation. Highly polysemous verbs are more complicated and may be involved in certain semantic structures. Semi-automatic sense grouping may be helpful for multimlinguital information retrieveal.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.1
• /
• pp.364-380
• /
• 2016

The most popular approach in object classification is based on the bag of visual-words model, which has several fundamental problems that restricting the performance of this method, such as low time efficiency, the synonym and polysemy of visual words, and the lack of spatial information between visual words. In view of this, an object classification based on weakly supervised E2LSH and saliency map weighting is proposed. Firstly, E2LSH (Exact Euclidean Locality Sensitive Hashing) is employed to generate a group of weakly randomized visual dictionary by clustering SIFT features of the training dataset, and the selecting process of hash functions is effectively supervised inspired by the random forest ideas to reduce the randomcity of E2LSH. Secondly, graph-based visual saliency (GBVS) algorithm is applied to detect the saliency map of different images and weight the visual words according to the saliency prior. Finally, saliency map weighted visual language model is carried out to accomplish object classification. Experimental results datasets of Pascal 2007 and Caltech-256 indicate that the distinguishability of objects is effectively improved and our method is superior to the state-of-the-art object classification methods.

• Proceedings of the Korean Society for Language and Information Conference
• /
• 2007.11a
• /
• pp.303-310
• /
• 2007

Named Entity Recognition (NER) is always limited by its lower recall resulting from the asymmetric data distribution where the NONE class dominates the entity classes. This paper presents an approach that exploits non-local information to improve the NER recall. Several kinds of non-local features encoding entity token occurrence, entity boundary and entity class are explored under Conditional Random Fields (CRFs) framework. Experiments on SIGHAN 2006 MSRA (CityU) corpus indicate that non-local features can effectively enhance the recall of the state-of-the-art NER systems. Incorporating the non-local features into the NER systems using local features alone, our best system achieves a 23.56% (25.26%) relative error reduction on the recall and 17.10% (11.36%) relative error reduction on the F1 score; the improved F1 score 89.38% (90.09%) is significantly superior to the best NER system with F1 of 86.51% (89.03%) participated in the closed track.

• Science of Emotion and Sensibility
• /
• v.22 no.4
• /
• pp.65-74
• /
• 2019

Extant research has proposed that the Object-Spatial-Verbal cognitive style can elucidate individual differences in the preference for modality-specific information. However, no studies have yet ascertained whether this type of information processing evinces structural correlations in the brain. Therefore, the current study used voxel-based morphometry (VBM) analyses to investigate individual differences in gray matter volumes based on the Object-Spatial-Verbal cognitive style. For this purpose, ninety healthy young adults were recruited to participate in the study. They were administered the Korean version of the Object-Spatial-Verbal cognitive style questionnaire, and their anatomical brain images were scanned. The VBM results demonstrated that the participants' verbal scores were positively correlated with regional gray matter volumes (rGMVs) in the right superior temporal sulcus/superior temporal gyrus, the bilateral parahippocampal gyrus/fusiform gyrus, and the left inferior temporal gyrus. In addition, the rGMVs in these regions were negatively correlated with the relative spatial preference scores obtained by individual participants. The findings of the investigation provide anatomical evidence that the verbal cognitive style could be decidedly relevant to higher-level language processing, but not to basic language processing.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.9-15
• /
• 2021

기계번역 품질 예측 (Quality Estimation, QE)은 정답 문장에 대한 참조없이 소스 문장과 기계번역 결과를 통해 기계번역 결과에 대한 품질을 수준별 주석으로 나타내주는 태스크이며, 다양한 활용도가 있다는 점에서 꾸준히 연구가 수행되고 있다. 그러나 QE 모델 학습을 위한 데이터 구성 시 기계번역 결과에 대해 번역 전문가가 교정한 문장이 필요한데, 이를 제작하는 과정에서 상당한 인건비와 시간 비용이 발생하는 한계가 있다. 본 논문에서는 번역 전문가 없이 병렬 또는 단일 말뭉치와 기계번역기만을 활용하여 자동화된 방식으로 한국어-영어 합성 QE 데이터를 구축하며, 최초로 단어 수준의 한국어-영어 기계번역 결과 품질 예측 모델을 제작하였다. QE 모델 제작 시에는 Cross-lingual language model (XLM), XLM-RoBERTa (XLM-R), multilingual BART (mBART)와 같은 다언어모델들을 활용하여 비교 실험을 수행했다. 또한 기계번역 결과에 대한 품질 예측의 객관성을 검증하고자 구글, 아마존, 마이크로소프트, 시스트란의 번역기를 활용하여 모델 평가를 진행했다. 실험 결과 XLM-R을 활용하여 미세조정학습한 QE 모델이 가장 좋은 성능을 보였으며, 품질 예측의 객관성을 확보함으로써 QE의 다양한 장점들을 한국어-영어 기계번역에서도 활용할 수 있도록 했다.

• Applied Microscopy
• /
• v.45 no.2
• /
• pp.101-105
• /
• 2015

Numerous digital image analysis techniques have been developed with regard to transmission electron microscopy (TEM) with the help of programming. DigitalMicrograph (DM, Gatan Inc., USA), which is installed on most TEMs as operational software, includes a script language to develop customized software for image analysis. Based on the DM script language, this work provides a script source listing for quantitative strain measurements based on a geometric phase analysis.

• Proceedings of the KIEE Conference
• /
• 1988.07a
• /
• pp.659-661
• /
• 1988

SPM is MDC Processor for string pattern expressed in directional chain code. In this paper we consider the string pattern matching algorithm (Leve-nstein Algorithm) whitch is portion of Dynamic Programing, and propose architecture of SPM and simulate it on the R-T level to evaluate its architecture. We used the C language as the hardware description language, and developed it on the IBM PC/AT Zenix system V OS environment.

• Annual Conference on Human and Language Technology
• /
• 1997.10a
• /
• pp.219-226
• /
• 1997

에이전트는 분산 환경에서 작업을 수행하는 지적인 특성을 갖는 응용 프로그램으로 정의되며, 연구 분야에 따라 다양한 의미로 해석이 가능하다. 그중에서 로봇 에이전트는 전세계에 산재된 방대한 양의 정보를 스스로 추적하며 새로운 정보를 찾는다. 로봇 에이전트에 대한 기존의 연구는 대부분 통계적인 목적이나 검색엔진을 위한 데이터의 수집을 목적으로 사용되었다. 많은 정보를 수집하기 위해 더 높은 성능의 로봇 에이전트들이 제작되었고, 이러한 프로그램들이 팽창하면서 네트워크를 과부하시키는 현상을 초래하게 되었다. 재귀적인 방법으로 수행되는 로봇 에이전트의 사용을 억제하기 위한 연구들이 많이 발표되었으나, 수동적인 방법에 의존하는 연구가 대부분이며 대표적인 것이 로봇 배제를 위한 표준안 정도이다. 본 연구에서는 이러한 로봇 에이전트의 문제점을 개선하여, 서버와 클라이언트간에 대화형으로 수행되는 인덱스 로봇 에이전트를 제안하며, 사용자의 요구에 따라 수행되는 로봇 에이전트에 의한 정보 획득의 방법을 시도하여 네트워크의 과부하를 억제하면서도 정보의 신뢰성과 정확성을 보장한다.

• Proceedings of the Korean Society for Language and Information Conference
• /
• 2002.02a
• /
• pp.271-280
• /
• 2002

A system to assist call routing task for telephone operators at the Directorate General of Telecommunications (DGT) in Taiwan is reported in this paper. The system was developed based on DGT organization profile with description of its six divisions instead of a corpus of recorded and transcribed call-routing dialogs. An acoustic module and an information retrieval module were built specifically for this task. The construction of IR module was based on term extraction and thesaurus discovery processes. By integrating acoustic and IR module, the system achieves satisfactory performance and provides a promising approach to call routing. Simulation results indicated that the proposed algorithm outperforms standard classification methods. A working system based on the proposed approach has been implemented and experimental results are presented.