• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.6
• /
• pp.663-667
• /
• 2010

A simple approach to semantic document-retrieval is to measure document similarity based on the bag-of-words representation, e.g., cosine similarity between two document vectors. However, such a syntactic method hardly considers the semantic similarity between documents, often producing semantically-unsound search results. We circumvent such a problem by combining supervised machine learning techniques with ontology information based on Markov logic. Specifically, Markov logic networks are learned from similarity-tagged documents with an ontology representing the diverse relationship among words. The learned Markov logic networks, the ontology, and the training documents are applied to the semantic document-retrieval task by inferring similarities between a query document and the training documents. Through experimental evaluation on real world question-answering data, the proposed method has been shown to outperform the simple cosine similarity-based approach in terms of retrieval accuracy.

• Journal of KIISE
• /
• v.45 no.1
• /
• pp.30-35
• /
• 2018

GuessWhat?! is a game in which two machine players, composed of questioner and answerer, ask and answer yes-no-N/A questions about the object hidden for the answerer in the image, and the questioner chooses the correct object. GuessWhat?! has received much attention in the field of deep learning and artificial intelligence as a testbed for cutting-edge research on the interplay of computer vision and dialogue systems. In this study, we discuss the objective function and characteristics of the GuessWhat?! game. In addition, we propose a simple solver for GuessWhat?! using a simple rule-based algorithm. Although a human needs four or five questions on average to solve this problem, the proposed method outperforms state-of-the-art deep learning methods using only two questions, and exceeds human performance using five questions.

• KIPS Transactions on Software and Data Engineering
• /
• v.6 no.4
• /
• pp.193-202
• /
• 2017

Development of Intelligent Systems involves effective integration of large-scaled knowledge processing and understanding, human-machine interaction, and intelligent services. Especially, in our project for development of a self-growing knowledge-based system with inference methodologies utilizing the big data technology, we are building a platform called WiseKB as the central knowledge base for storing massive amount of knowledge and enabling question-answering by inferences. WiseKB thus requires an effective methodology to analyze diverse requirements convoluted with the integration of various components of knowledge representation, resource management, knowledge storing, complex hybrid inference, and knowledge learning, In this paper, we propose an integrated requirement analysis method that blends the traditional sequential method and the iterative-incremental method to achieve an efficient requirement analysis for large-scale systems.

• The KIPS Transactions:PartD
• /
• v.18D no.5
• /
• pp.329-338
• /
• 2011

Terminology recognition system which is a preceding research for text mining, information extraction, information retrieval, semantic web, and question-answering has been intensively studied in limited range of domains, especially in bio-medical domain. We propose a domain independent terminology recognition system based on machine learning method using dictionary, syntactic features, and Web search results, since the previous works revealed limitation on applying their approaches to general domain because their resources were domain specific. We achieved F-score 80.8 and 6.5% improvement after comparing the proposed approach with the related approach, C-value, which has been widely used and is based on local domain frequencies. In the second experiment with various combinations of unithood features, the method combined with NGD(Normalized Google Distance) showed the best performance of 81.8 on F-score. We applied three machine learning methods such as Logistic regression, C4.5, and SVMs, and got the best score from the decision tree method, C4.5.

• Hwankyungkyoyuk
• /
• v.23 no.2
• /
• pp.82-96
• /
• 2010

The purposes of this study are to investigate the perception of renewable energy technology among elementary school teachers, and confirm whether elementary school teachers have basic knowledge about renewable energy sources, including solar, wind, and tidal power generation. We conducted preliminary interviews to gather information related to other studies about renewable energy. We developed the last interview question about the perception and knowledge of elementary school teachers regarding renewable energy. This study analyzed the transcribed responses of 10 elementary school teachers in Siheung-city, Gyeonggi-do, following 30-minute interviews. The study's findings are as follows. First, elementary school teachers recognize that they are unfamiliar with concepts and they have only shallow content knowledge about renewable energy. And they tended to distorted to other concepts, and analyze to different meanings. Second, elementary school teachers thought that knowledge about renewable energy should be part of a well-rounded education. And they felt positively about solar energy and wind power energy generation but they had a negative view towards tidal power generation because it destroys tideland. Third, teachers tended to confuse solar heat energy and geothermal energy, they tend to think this two energy sources the same. Teachers had generally correct concepts about wind power energy generation. In the case of tidal power generation, elementary school teachers answered mechanically that it is possible on the western sea, and that 'the difference between the rise and fall of the tide' grows. But they could not talk in depth about 'the difference between the rise and fall of the tide' and the force of waves. This suggests that they are answering by simple memorization and without deep understanding.

• Annual Conference on Human and Language Technology
• /
• 2006.10e
• /
• pp.268-272
• /
• 2006

질의응답 시스템은 사용자 질의에 해당하는 정답을 찾기 위해서 세부 분류된 개체명을 사용한다. 이러한 세부 분류 개체명 인식을 위해서 대부분의 시스템이 일반 대분류 개체명인식 후에 사전 등을 이용하여 세부 분류로 나누는 방법을 이용하고 있다. 본 논문에서는 질의응답 시스템을 위한 세부 분류 개체명 인식을 위해서 Conditional Random Fields를 이용한다. 개체명 인식의 과정을 개체명 경계 인식과 경계가 인식된 개체명의 클래스 분류의 두 단계로 나누어, 개체명 경계 인식에 Conditional Random Fields를 이용하고, 경계 인식된 개체명의 클래스 분류에는 Maximum Entropy를 이용한다. 실험결과 147개의 세부분류 개체명 인식에 대해서 정확도 85.8%, 재현률 81.1%. F1=83.4의 성능을 얻었고. baseline model 보다 학습 시간이 27%로 줄고 성능은 증가하였다. 또한 제안된 세부 분류개체명 인식기를 이용하여 질의응답 시스템에 적용한 결과 26%의 성능향상을 보였다.

• Journal of the Korean Society for information Management
• /
• v.28 no.1
• /
• pp.89-104
• /
• 2011

Large-scaled information extraction plays an important role in advanced information retrieval as well as question answering and summarization. Information extraction can be defined as a process of converting unstructured documents into formalized, tabular information, which consists of named-entity recognition, terminology extraction, coreference resolution and relation extraction. Since all the elementary technologies have been studied independently so far, it is not trivial to integrate all the necessary processes of information extraction due to the diversity of their input/output formation approaches and operating environments. As a result, it is difficult to handle scientific documents to extract both named-entities and technical terms at once. In order to extract these entities automatically from scientific documents at once, we developed a framework for scientific core entity extraction which embraces all the pivotal language processors, named-entity recognizer and terminology extractor.

• KIPS Transactions on Software and Data Engineering
• /
• v.5 no.11
• /
• pp.535-540
• /
• 2016

Coreference resolution is to identify mentions in documents and is to group co-referred mentions in the documents. It is an essential step for natural language processing applications such as information extraction, event tracking, and question-answering. Recently, various coreference resolution models based on ML (machine learning) have been proposed, As well-known, these ML-based models need large training data that are manually annotated with coreferred mention tags. Unfortunately, we cannot find usable open data for learning ML-based models in Korean. Therefore, we propose an efficient coreference resolution model that needs less training data than other ML-based models. The proposed model identifies co-referred mentions using random forests based on sieve-guided features. In the experiments with baseball news articles, the proposed model showed a better CoNLL F1-score of 0.6678 than other ML-based models.

• Journal of KIISE:Software and Applications
• /
• v.33 no.6
• /
• pp.564-573
• /
• 2006

Temporal information plays an important role in natural language processing (NLP) applications such as information extraction, discourse analysis, automatic summarization, and question-answering. In the topic detection and tracking (TDT) area, the temporal information often used is the publication date of a message, which is readily available but limited in its usefulness. We developed a relatively simple NLP method of extracting temporal information from Korean news articles, with the goal of improving performance of TDT tasks. To extract temporal information, we make use of finite state automata and a lexicon containing time-revealing vocabulary. Extracted information is converted into a canonicalized representation of a time point or a time duration. We first evaluated the extraction and canonicalization methods for their accuracy and investigated on the extent to which temporal information extracted as such can help TDT tasks. The experimental results show that time information extracted from text indeed helps improve both precision and recall significantly.

• International Journal of Advanced Culture Technology
• /
• v.9 no.4
• /
• pp.180-187
• /
• 2021

This thesis explores the place and role of a performer's action from a perspective of a director and playwright David Mamet's concept for performer training. This thesis takes inspiration from the idea of Mamet's simple and practical investigation specifically in text-based approach with a performer's bodily function on stage. For Mamet, the writings and practices of many different body-centered training are not rooted in the principle and nature of acting/performance. Reconsidering complicated approaches particularly psychological-oriented theory, practice, and assumption draw on several practitioners takes us beyond the field of visible and/or outer appearance of a performer which in turn leads the performer's body to be as abstract therefore not to being in the moment on stage. Arming the points, we argue that whatever disciplines and/or methods necessarily need to meet the principles and demands of acting/performance/theatre to connect to the materials, an action/objective given by a specific playwright which the performer must inhabit through his/her body. Out of the context, any 'method' serves no purpose. That is, the mechanics of an action is an extension of addressing what a performer's specific needs which shifts his/her body to respond appropriately to the theatrical demands. Taking this argument further, we claim that the purpose of performer training should not be understood as learning and improving techniques or skills for his/her self-perfection. The research finding shows that this resembles to the phenomenon that the visible very often precedes the invisible where the performer's body lose a clarity with no more chance to happen and/or change the event(s). Rather, it is a process of learning what/how to learn which in turn brings us back to the central question of why we do training for what purpose in this contemporary era. Exploring and answering these questions is not only a way to employ the key materials applicable to the theatrical demands but also to achieve the identify as a professional performer/doer on stage.