구문 의존 경로에 기반한 단백질의 세포 내 위치 인식

김미영 한국정보처리학회 2008 정보처리학회논문지. 소프트웨어 및 데이터 공학 Vol.15 No.4

A protein’s subcellular localization is considered an essential part of the description of its associated biomolecular phenomena. As the volume of biomolecular reports has increased, there has been a great deal of research on text mining to detect protein subcellular localization information in documents. It has been argued that linguistic information, especially syntactic information, is useful for identifying the subcellular localizations of proteins of interest. However, previous systems for detecting protein subcellular localization information used only shallow syntactic parsers, and showed poor performance. Thus, there remains a need to use a full syntactic parser and to apply deep linguistic knowledge to the analysis of text for protein subcellular localization information. In addition, we have attempted to use semantic information from the WordNet thesaurus. To improve performance in detecting protein subcellular localization information, this paper proposes a three-step method based on a full syntactic dependency parser and WordNet thesaurus. In the first step, we constructed syntactic dependency paths from each protein to its location candidate, and then converted the syntactic dependency paths into dependency trees. In the second step, we retrieved root information of the syntactic dependency trees. In the final step, we extracted syn-semantic patterns of protein subtrees and location subtrees. From the root and subtree nodes, we extracted syntactic category and syntactic direction as syntactic information, and synset offset of the WordNet thesaurus as semantic information. According to the root information and syn-semantic patterns of subtrees from the training data, we extracted (protein, localization) pairs from the test sentences. Even with no biomolecular knowledge, our method showed reasonable performance in experimental results using Medline abstract data. Our proposed method gave an F-measure of 74.53% for training data and 58.90% for test data, significantly outperforming previous methods, by 12-25%. 단백질의 세포 내 위치를 인식하는 것은 생물학 현상의 기술에 있어서 필수적이다. 생물학 문서의 양이 늘어남에 따라, 단백질의 세포 내 위치 정보를 문서 내용으로부터 얻기 위한 연구들이 많이 이루어졌다. 기존의 논문들은 문장의 구문 정보를 이용하여 정보를 얻고자 하였으며, 언어학적 정보가 단백질의 세포 내 위치를 인식하는 데 유용하다고 주장하고 있다. 그러나, 이전의 시스템들은 구문 정보를 얻기 위해 부분 구문분석기만을 사용하였고 재현율이 좋지 못했다. 그러므로 단백질의 세포 내 위치 정보를 얻기 위해 전체 구문분석기를 사용할 필요가 있다. 또한, 더 많은 언어학적 정보를 위해 의미 정보 또한 사용이 가능하다. 단백질의 세포 내 위치 정보를 인식하는 성능을 향상시키기 위하여, 본 논문은 전체 구문분석기와 어휘망(WordNet)을 기반으로 한 방법을 제안한다. 첫 번째 단계에서, 각 단백질 단어로부터 그 단백질의 위치후보에까지 이르는 구문 의존 경로를 구축한다. 두 번째 단계에서, 구문 의존 경로의 루트 정보를 추출한다. 마지막으로, 단백질 부분트리와 위치 부분트리의 구문-의미 패턴을 추출한다. 구문 의존 경로의 루트와 부분트리로부터 구문태그와 구문방향을 구문 정보로서 추출하고, 각 노드 단어의 의미태그를 의미 정보로서 추출한다. 의미태그로는 어휘망의 동의어 집합(synset)을 사용한다. 학습데이터에서 추출한 루트 정보와 부분트리의 구문-의미 패턴에 따라서, 실험데이터에서 (단백질, 위치) 쌍들을 추출했다. 어떤 생물학적 지식 없이, 본 논문의 방법은 메드라인(Medline) 요약 데이터를 사용한 실험 결과에서 학습데이터에 대해 74.53%의 조화평균(F-measure), 실험데이터에 대해서는 58.90%의 조화평균을 보였다. 이 실험은 기존의 방법들보다 12-25%의 성능향상을 보였다.

ERP실험을 통한 EFL 화자의 통사적 민감성 연구

김성헌,배상희,박명관,정원일 한국생성문법학회 2013 생성문법연구 Vol.23 No.4

This paper explores the syntactic sensitivity by EFL speakers through the ERP experiment, paying special attention to the syntactic reanalysis for the awareness on phases in the language processing within the phase-based structures (i.e., CP, vP), which include transitive verbs or intransitive verbs. Through ERP (event-related potentials), we investigate how EFL speakers respond to syntactic errors within a sentence and how sensitive they are to the syntactic information. Clahsen and Felser (2006a, b) claim that, based on the SSH (Shallow Structure Hypothesis), L2 learners who have learned their L2 after acquiring their native language process the L2 differently from native speakers. Zawiszewski, Gutiérrez, Fernández, and Laka (2011), however, report that L2 speakers show the language process as analogous to that of L1 speakers if the proficiency of L2 speakers is close to that of L1 speakers. On the basis of these studies, we examined whether Korean EFL speakers use the syntactic information and are sensitive to syntactic structures as in the language processing of L1 speakers and other English L2 speakers. To achieve the aim of this study, we try to explain the results by examining their ERP components and brain maps. Our results suggest that EFL speakers show LAN (left anterior negativity) in the comparison of transitive sentences as in Kim and Sikos (2011), unlike L1 and L2 speakers who record a P600 during the language process using structured syntactic information. Thus, this study verifies that EFL speakers also try to use syntactic information for processing the input phase information and are syntactically sensitive to the properties of verbs.

문법 교육 : 한국어 학습자를 위한 어휘 정보: {소리}

신현숙 ( Hyon Sook Shin ) 한국문법교육학회 2010 문법 교육 Vol.13 No.-

Korean noun {소리 soli [(a) sound/ a voice/ a ballad]} belongs to the basic lexical items in the Korean language. It will be meaningful to discuss its lexical information for Korean language learners and Korean language computer program developers. It is also important to examine how Korean language users use the item in the appropriate contexts and how they extend its basic form and meaning. For this purpose, we categorized the usage and contexts of {soli} in real data base on the item`s syntactic construction methods. Therefore we would like to find the characteristics of {X} and {Y} within the frame: {X}+{soli}+{Y}}. We also analyzed its combinational constituents based on the criteria of the syntactic information and semantic information. We also discussed the semantic information, especially meaning extension of {soli} based upon the five meaning of the major dictionary in Korea. Korean language learners will be able to learn the usage and construction of {soli} in system with this kind of lexical information in which we have suggested in this paper. Furthermore our information also applies to the Korean language computer program. For instance, we can apply this system when analyzing the Korean language construction as a tool.

The Prosody-semantic Interface in the Resolution of L2 Syntactic Ambiguity

Jeonghwa Shin 한국영어학학회 2019 영어학연구 Vol.25 No.3

The present study examines the interface between semantic and prosodic boundary information in the resolution of L2 syntactic ambiguity. In a cross-modal auditory comprehension experiment, 5 native English speakers as a control group and 20 Korean learners of English listened to syntactically ambiguous English sentences with or without semantically and prosodically-biased information. The results showed that both native English speakers and L2 learners utilized semantic and prosodic information to resolve the syntactic ambiguities. Different types of semantic information yielded different effects on the use of prosodic boundary cues in both groups. The native and non-native groups, however, were different in their preferences for the use of semantic and prosodic cues, especially when the two cues were in conflict. Native English listeners relied more on semantic cues than on prosodic cues to resolve the syntactic ambiguity when the two cues were conflicting, and they utilized prosodic information more reliably only when the syntactic ambiguity was not resolved with semantic information. Whereas, L2 learners’ parsing was influenced by the prosodic cues even when the semantic cues clearly directed the meaning of the ambiguous sentence. The results indicate that L2 sentence processing can be accounted for by the universal interactive-constraint model (MacDonald et al. 1994). However, L2 auditory sentence processing is different from L1 in that L2 learners are more susceptible to cognitive interference from conflicting cues than native speakers (Cunnings 2016).

Transfer of L1 Semantic and Syntactic Information in the Acquisition of L2 Transitive Verbs

( Bokyoung Jung ),( Junkyu Lee ) 서강대학교 언어정보연구소 2016 언어와 정보 사회 Vol.27 No.-

The purpose of this study is to explore how the activated L1 lemma information (the semantic and syntactic information) is transferred to influence L2 lexical use (Jiang 2004). This study accounts for the mediation of L1 lemma information happening to three low-level middle school students compared to adult L2 acquisition process (Jiang 2000; Jiang 2004). To examine the mediation between L1 lemma information and L2 lexical form, the semantic mapping of seventy transitive verbs was associated with L1 lemma information. The analyzed qualitative data suggested that the semantic and syntactic information of L1 lemma supported understanding lexical features of transitive verbs. As a pedagogical implication, the continuous practice in mapping promoted the spontaneous form-meaning connection between L1 and L2.

Transfer of L1 Semantic and Syntactic Information in the Acquisition of L2 Transitive Verbs

정보경,이준규 서강대학교 언어정보연구소 2016 언어와 정보 사회 Vol.27 No.-

The purpose of this study is to explore how the activated L1 lemma information (the semantic and syntactic information) is transferred to influence L2 lexical use (Jiang 2004). This study accounts for the mediation of L1 lemma information happening to three low-level middle school students compared to adult L2 acquisition process (Jiang 2000; Jiang 2004). To examine the mediation between L1 lemma information and L2 lexical form, the semantic mapping of seventy transitive verbs was associated with L1 lemma information. The analyzed qualitative data suggested that the semantic and syntactic information of L1 lemma supported understanding lexical features of transitive verbs. As a pedagogical implication, the continuous practice in mapping promoted the spontaneous form-meaning connection between L1 and L2.

한국어 단어 재인 시 선행 어절의 의미적‧통사적 정보가 미치는 영향

남기춘(Nam, Ki-chun),이은하(Lee, Eun-ha),전현애(Jeon, Hyeon-ae),이선영(Lee, Sun-young),김영주(Kim, Young-joo) 국어국문학회 2018 국어국문학 Vol.0 No.184

본 연구는 한국어 모어 화자가 서로 인접한 어절과 단어로 이뤄진 단일 구를 처리할 때 선행어절의 의미적‧통사적 맥락이 후행단어의 재인에 영향을 미치는지 그리고 의미적‧통사적 정보는 각기 어떤 순서로 처리되는지를 고찰했다. 이를 위해 형용사 어간과 어미 ‘-ㄴ/은’이 결합된 어절을 점화로 그리고 일반명사 단어를 표적으로 하는 점화 어휘판단 과제를 사용했다. 연구결과에 따르면, 점화어절과 표적단어가 의미적‧통사적으로 호응하는 조건은 의미적으로 상통하지 않지만 통사적으로는 호응하는 조건이나 의미적‧통사적으로 모두 상통하지 않는 조건보다 빠른 반응시간을 나타냈다. 아울러 의미적으로 상통하지 않지만 통사적으로는 호응하는 조건에서 의미적‧통사적으로 모두 상통하지 않는 조건보다 표적단어가 더 빨리 처리되는 양상을 보였다. 이는 형용사 어절과 명사로 이루어진 명사구 내에서도 의미적 점화가 일어나며, 점화어절의 의미적 정보를 바탕으로 어휘적 후보를 활성화할 수 없는 환경에서도 통사적 정보가 후속단어 처리에 독자적 영향력을 발휘함을 시사한다. The present study aims at investigating how syntactic and semantic context constraints affect word recognition in Korean. Participants performed a primed lexical decision task where a target noun was preceded by a semantically and syntactically congruent adjective eogeol, a syntactically congruent but semantically incongruent adjective eojeol, or a semantically and syntactically incongruent adjective eojeol. The results revealed that the syntactically and semantically congruent context reduced significantly reaction time for the subsequent target noun compared with the syntactically congruent but semantically incongruent context. More importantly, the syntactically congruent but semantically incongruent condition produced faster response times than semantically and syntactically incongruent condition. These findings provide supportive evidences for the possibility of both semantic and syntactic priming effects in a minimal phrase structure and the assumption of the serial model claiming that syntactic information may be processed earlier than semantic information.

의미적, 통사적 요소의 언어 간 전이

정보경(Jung Bo kyoung),이준규(Lee Jun kyu) 한국외국어대학교 영미연구소 2016 영미연구 Vol.37 No.-

본 연구는 모국어 정보에서, 특히 의미적, 통사적 요소를 영어의 용법으로 전이하는 과정에 초점이 있다. 학습자들의 모국어 정보가 영어 동사에 입력되면서, 제2외국어의 형식이 학습자들의 모국어 어휘에서 전이된 의미적, 통사적 정보와 결합하는 현 상이 발생한다. 한국어와 영어 동사 간 형식-의미 구조 형성과정을 알아보기 위해, 영어 어휘의 모국어 번역을 구조도 활동 중에 학습자들에게 제시했다. 그에 따라, 학습자들은 모국어 번역을 쉽게 받아들이고 번역을 이용하여 다른 종류의 영어 동사와 관련된 모국어 지식을 활성화시켰다. 구조도 활동의 결과를 통하여 모국어 의미적, 통사적 정보 전이과정을 제시할 수 있었다. 실험 참여자들은 동사와 보어의 문장 안에서 의미 역할을 이해하기 위해 모국어 정보를 활용할 수 있었다. 하지만 문장 안에서 보어의 위치를 이해하고 모국어 정보를 영어의 용법에 맞게 전이하는데 시간이 소요되었다. 정보를 전이하는 과정에서, 모국어 정보가 잘못된 영향을 끼치기도 했다. 사후 시험결과, 학습자들은 대부분의 경우 모국어 정보를 동사의 의미적, 통사적 요소와 연결할 수 있었다. 교육학적 영향으로, 올바른 수정과 피드백은 모국어와 제 2외국어 형식-의미 구조 형성을 개선한다고 말할 수 있다. This study investigates the transfer of L1 information, especially semantic and syntactic information, into English specific usage. As learners’ L1 information is mapped onto English verbs, L2 forms are integrated with the semantic and syntacticinformation transferred from L1 lexical entries. To examine into the form-meaning mapping between Korean and English verbs, L1 translations corresponding to L2 words were provided during the mapping activities. Accordingly, learners easily adopted L1 translations and activated their L1 information to link with different types of English verbs. The outcome of the mapping activities might account for the process of transferring the semantic and syntactic information of L1. Participants benefited from their L1 information to understand the semantic role of verbs and their complements. The result of the post-tests suggested that learners mostly could connect L1 information with the semantic and syntactic feature of verbs. As a pedagogical implication, the corrective specification and feedback enhanced the form-meaning mapping between L1 and L2.

조사(助詞) ‘-은/는’과 구문구조(構文構造)

오충연(吳充淵) ( Oh¸ Choong-yeun ) 한국어문교육연구회 2021 어문연구(語文硏究) Vol.49 No.3

조사 ‘-은/는’은 다른 보조조사와 달리 어휘로 대체될 만한 의미가 없고 격조사 ‘-이/가’, ‘-을/를’과 대립하여 共起制約을 보인다. 본 연구에서는 그 원인이 주격과 목적격의 구성이 기저에서 화제와 초점의 구성과 일치하기 때문으로 분석하였다. 그러므로 별도의 정보 표지가 없이 구현된 無標的인 문장에서의 화제와 초점은 곧 주어와 목적어이다. 국어에서는 전달 층위의 형태론적 장치에 의해서 이러한 정보 구조가 재구성되는데, ‘-은/는’이 여기에 적용되는 형식이다. 그러나 기존에 알려진 바처럼 ‘-은/는’이 화제나 초점의 정보 기능을 하거나 대조의 의미를 가지게 되는 것은 적용되는 논항의 위치에 의해 결정되는 것이므로, ‘-은/는’ 자체의 순수한 기능은 ‘정보구조의 강화’라는 一元的인 原理로 설명하였다. 이에 따라 ‘-은/는’의 구조적 위치가 정보구조를 어떻게 변환시키는지를 고찰하였다. 한편 ‘-은/는’의 기능이 구조격과 교차하는 현상을 통시적 과정에 의해 발생하는 것으로 推定하였다. The Korean post-positional particle -은/는 has been defined as an auxiliary particle, but it differs from other auxiliary particles in that it has no lexical meaning and has co-occurrence restrictions with -이/가 and -을/를 as structural case markers. This paper analyzes these restrictions as due to an overlap with the syntactic function of subjective or objective. The structure of the sentence is differentiated from the propositional structure, and the subject and object cases are structural cases assigned at different levels from the thematic structure of the verb of the sentence. In particular, the point of view of the speaker determines which of the participants is adopted as the subject, which depends on the importance of the respective information. The objective case is also organized based on the construction of the subjective argument. Therefore, in a default sentence with a basic information structure, the subjective is the topic and the objective is the focus. In a Korean sentence, the basic information structure of the sentence is transformed to the surface layer by the word order, which reinforces the degree of information as topic or focus, depending on whether -은/는 appears with the default word order. When it is added to an argument that is not the focus, it serves to set the scope of the topic of the entire sentence to the argument. When used with word order change, this changes the role of the arguments in the information structure.

영어 주어-동사 도치 구문 유발요인에 대한 코퍼스 기반 연구

서아리,이필환 한국영어학학회 2022 영어학연구 Vol.28 No.3

The purpose of this study is to investigate the causes of English Subject–Verb Inversion(SVI) based on a personally built corpus. This study also examines the characteristics of the verbs used in SVI. The corpus amounts to approximately 600,000 words of English in total. The findings of the study are as follows. First, this study revealed that both syntactic complexity and information status influence word order in English, in that most cases of SVI in the corpus were post-heavy and post-new, which means that an element of heavy weight and new information tends to be located toward the end of a sentence. Therefore, this study proves that SVI is in correlation with syntactic complexity and information status. It also identified that most of the verbs found in the SVI construction were be and intransitive verbs. In other words, the verbs used in the SVI construction are informationally light without representing new information. However, given that the corpus size of this study is small compared to other corpus-driven studies, future research requires increasing the data size of the corpus. Moreover, this study only focuses on written texts, so it is necessary to include spoken texts for a comparative analysis of written and spoken texts.