• ETRI Journal
• /
• v.39 no.5
• /
• pp.652-661
• /
• 2017

A mention has a noun or noun phrase as its head and constructs a chunk that defines any meaning, including a modifier. Mention detection refers to the extraction of mentions from a document. In mentions, coreference resolution refers to determining any mentions that have the same meaning. Pointer networks, which are models based on a recurrent neural network encoder-decoder, outputs a list of elements corresponding to an input sequence. In this paper, we propose mention detection using pointer networks. This approach can solve the problem of overlapped mention detection, which cannot be solved by a sequence labeling approach. The experimental results show that the performance of the proposed mention detection approach is F1 of 80.75%, which is 8% higher than rule-based mention detection, and the performance of the coreference resolution has a CoNLL F1 of 56.67% (mention boundary), which is 7.68% higher than coreference resolution using rule-based mention detection.

• Journal of KIISE
• /
• v.44 no.8
• /
• pp.774-781
• /
• 2017

Mention detection systems use nouns or noun phrases as a head and construct a chunk of text that defines any meaning, including a modifier. The term "mention detection" relates to the extraction of mentions in a document. In the mentions, a coreference resolution pertains to finding out if various mentions have the same meaning to each other. A pointer network is a model based on a recurrent neural network (RNN) encoder-decoder, and outputs a list of elements that correspond to input sequence. In this paper, we propose the use of mention detection using pointer networks. Our proposed model can solve the problem of overlapped mention detection, an issue that could not be solved by sequence labeling when applying the pointer network to the mention detection. As a result of this experiment, performance of the proposed mention detection model showed an F1 of 80.07%, a 7.65%p higher than rule-based mention detection; a co-reference resolution performance using this mention detection model showed a CoNLL F1 of 52.67% (mention boundary), and a CoNLL F1 of 60.11% (head boundary) that is high, 7.68%p, or 1.5%p more than coreference resolution using rule-based mention detection.

• KIISE Transactions on Computing Practices
• /
• v.21 no.4
• /
• pp.333-337
• /
• 2015

In this paper, we suggest a Coreference Resolution system for Korean using Mention Pair with SVM. The system introduced in this paper, also be able to extract Mention from document which is including automatically tagged name entity information, dependency trees and POS tags. We also built a corpus, including 214 documents with Coreference tags, referencing online news and Wikipedia for training the system and testing the system's performance. The corpus had 14 documents from online news, along with 200 question-and-answer documents from Wikipedia. When we tested the system by corpus, the performance of the system was extracted by MUC-F1 55.68%, B-cube-F1 57.19%, and CEAFE-F1 61.75%.

• ETRI Journal
• /
• v.38 no.6
• /
• pp.1207-1217
• /
• 2016

The general method of machine learning has encountered disadvantages in terms of the significant amount of time and effort required for feature extraction and engineering in natural language processing. However, in recent years, these disadvantages have been solved using deep learning. In this paper, we propose a mention pair (MP) model using deep learning, and a system that combines both rule-based and deep learning-based systems using a guided MP as a coreference resolution, which is an information extraction technique. Our experiment results confirm that the proposed deep-learning based coreference resolution system achieves a better level of performance than rule- and statistics-based systems applied separately

• Journal of Service Research and Studies
• /
• v.10 no.2
• /
• pp.59-76
• /
• 2020

The purpose of this study is to analyze the impacts of the lasciviousness of news anchor's mention on viewers'negative emotions, news reliability and their attitude towards the broadcasting company. This study also analyzed the moderating effect of anchor's attractiveness and viewers' sex. First, the more lascivious anchor's mention in news report gets, the more negative the viewers felt. Second, stronger lascivious expressions of news anchor's mention in news report had a negative effect on news reliability. Third, the moderating effect of the anchor attractiveness was found when news anchor's mention influences the viewers' attitude towards the broadcasting company : those who thought news anchor attractive showed less negative emotions and their news reliability and attitude towards the broadcasting company were higher compared to those who thought news anchor less attractive. Fourth, the moderating effect of the viewers' sex was found when news anchor's mention influences on viewers' negative emotions and the viewers' attitude towards the broadcasting company. This study has an academic and practical implication by studying the lasciviousness of news anchor's mention and anchor's attractiveness. This study is also a new approach of integrating the fields of journalism : News report and Anchor into the marketing fields : Attractiveness and Reliability. This can be meaningful for both journalism and marketing field.

• Explosives and Blasting
• /
• v.16 no.3
• /
• pp.16-24
• /
• 1998

Methods of Rock fragmentation are used rock of housing repair development at KU-SAN DONG area in seoul Youn-Pyong Ku. So, Theorical analyses of the effect of vibration and frequency on structural damage around old housed also discussed. The results can be summarized as follows: 1. A area(Rock area not more than 15m Ku-San Mention) Some Empirical equations were obtained $V=K\{{\frac{D}{W}}1/3\}^{-n}$ where the values for n and K are estimated to be -1.64 and 94 respectively, this values were obtained only theorical analyses. If we have 125g charge this area is impossible blasting operation, so this area must be worked by SRS(Super Rock Splitter) method. 2. B area(Rock area from 15m to 25m in a boundary line from Ku-San Mention) This area charge is about 125g in a delay time by some empirical equation s. So, this area can be blasting operations by small charge. 3. C area(Rock area from 25m to 35m in a boundary line from Ku-San Mention) This area charge is about 500g in delay time by some empirical equation s. So, this area can be blasting operations by middle charge. 4. D area(Rock area more then 35m in a boundary line from Ku-San Mention) This area charge is about 1000g in a delay time by some empirical equation s. So, this area can be blasting operations by middle charge.

• Korean Journal of Child Studies
• /
• v.19 no.2
• /
• pp.185-200
• /
• 1998

This research examined the development in Korean children of the comprehension of complex sentences. The relative difficulty in comprehension of the temporal conjunctions "before" and "after" was investigated. The order of mention, contextual support, and syntactic appearance was controlled. The role of cognitive strategies and developmental changes in the comprehension of these conjunction was included in this study. Subjects were 90 preschool children between 3 and 5 years of age. The task was a sentence-picture matching problem having 3 types of sentences combining temporally with "before" or "after". The results were that developmental changes in comprehension of the temporal conjunctions "before' and "after" in Korean children depended on the order of mention, contextual support, and such syntactic factors as the position of the subject of the sentence. The importance of the consistency in the occurrence of events and the order of mention in the acquisition of complex sentences among Korean children is similar to the acquisition of complex sentences in other languages.

• Annual Conference on Human and Language Technology
• /
• 2015.10a
• /
• pp.224-227
• /
• 2015

상호참조해결은 특정 개체에 대해 다르게 표현한 단어들을 서로 연관지어 주며, 이러한 개체에 대해 표현한 단어들을 멘션(mention)이라 하며, 이런 멘션을 찾아내는 것을 멘션탐지(mention detection)라 한다. 멘션은 명사나 명사구를 기반으로 정의되며, 명사구의 경우에는 수식어를 포함하기 때문에 멘션탐지를 순차 데이터 문제(sequence labeling problem)로 정의할 수 있다. 순차 데이터 문제에는 Recurrent Neural Network(RNN) 종류의 모델을 적용할 수 있으며, 모델들은 Long Short-Term Memory(LSTM) RNN, LSTM Recurrent CRF(LSTM-CRF), Bidirectional LSTM-CRF(Bi-LSTM-CRF) 등이 있다. LSTM-RNN은 기존 RNN의 그레디언트 소멸 문제(vanishing gradient problem)를 해결하였으며, LSTM-CRF는 출력 결과에 의존성을 부여하여 순차 데이터 문제에 더욱 최적화 하였다. Bi-LSTM-CRF는 과거입력자질과 미래입력자질을 함께 학습하는 방법으로 최근에 가장 좋은 성능을 보이고 있다. 이에 따라, 본 논문에서는 멘션탐지에 Bi-LSTM-CRF를 적용할 것을 제안하며, 각 딥 러닝 모델들에 대한 비교실험을 보인다.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.20-23
• /
• 2020

개체 연결이란 문서에서 등장한 멘션(Mention)들을 지식 기반(Knowledge Base)상의 하나의 개체에 연결하는 문제를 말한다. 개체 연결은 개체를 찾는 멘션 탐지(mention detection)과정과 인식된 멘션에 대해 중의성을 해결하여 하나의 개체를 찾는 개체 중의성 해결(Entity disambiguation)과정으로 구성된다. 본 논문에서는 개체 정보를 강화하기 위해 wikipedia2vec정보를 결합하여 Entity 정보를 강화하고 문장 내에 모든 개체 정보를 활용하기 위해 집합적 개체를 정의하고 그래프 구조를 표현하기 위해 GNN을 활용하여 기존보다 높은 성능을 이끌어내었다.

• Korean Journal of Cognitive Science
• /
• v.21 no.2
• /
• pp.253-282
• /
• 2010

Two experiments were conducted to explore the effects of anaphora and demonstrative ellipsis on reference resolution. This study assumed that two type of ellipsis could be sensitive to antecedents' saliency: the reverse typicality and mention order of antecedents. The muti-task approach measured the antecedent's activation level and processing load for the conflict resolution of theories of anaphoric resolution. In Experiment 1, using ellipsis for anaphora, participants read a series of sentence pairs by self-paced and performed a probe recognition test. The results showed the main effects of antecedent's typicality and mention order in both tasks. In Experiment 2, using noun phrase without demonstrative for anaphora, participants read a series of sentence pairs by self-paced and performed a probe recognition test. The results showed main effects of mention order of antecedents for probe recognition task only. The first antecedent was recognized faster than the second one. The results of two experiments suggested that anaphora type and antecedent's saliency were dynamically interact in reference resolution for Korean.