간행물

한국언어정보학회> 언어와 정보

언어와 정보 update

LANGUAGE AND INFORMATION

  • : 한국언어정보학회
  • : 어문학분야  >  언어학
  • : KCI등재
  • :
  • : 연속간행물
  • : 연3회
  • : 1226-7430
  • :
  • :

수록정보
23권3호(2019) |수록논문 수 : 2
간행물 제목
23권3호(2019년) 수록논문
권호별 수록 논문
| | | |

KCI등재

1Leveraging More Fine-grained Representation to Reduce Instability within Word Embeddings

저자 : Suzi Park , Hyopil Shin

발행기관 : 한국언어정보학회 간행물 : 언어와 정보 23권 3호 발행 연도 : 2019 페이지 : pp. 1-18 (18 pages)

다운로드

(기관인증 필요)

초록보기

Suzi Park and Hyopil Shin. 2019. This paper investigates the notion of instability in word embeddings, primarily focused on the Korean language, by analyzing the effects of instability on several tasks using word level, character level, and sub-character level embeddings. Korean has a complex morphological system and is written at units smaller than that of a single character, making it optimal for comparing the three types of embeddings. We first carry out several intrinsic evaluations such as word similarity and analogy tests, as well as an extrinsic evaluation of the embeddings on the task of sentiment analysis. These results show that sub-character embeddings outperform the other two types, obtaining both higher and more consistent results, especially when the corpus size is large. Based on these findings, our hypothesis is that embeddings trained on smaller units and larger corpora would also be a better choice in terms of stability. To investigate this, we implemented two stability tests, looking into cosine similarity standard deviations and intersection of nearest neighbors among models. Our findings demonstrate that indeed, the smaller the embedding unit and the larger the corpus, the more stable the embedding type is. Importantly, our results illustrate the necessity of taking instability into consideration when using word embeddings within a given experiment.

KCI등재

2한국어 비교문장을 이용한 문장 진리치 판단에 관한 ERP 연구

저자 : 남윤주 ( Yunju Nam )

발행기관 : 한국언어정보학회 간행물 : 언어와 정보 23권 3호 발행 연도 : 2019 페이지 : pp. 19-38 (20 pages)

다운로드

(기관인증 필요)

초록보기

The present study investigated the underlying cognitive mechanism of the truth-value validation process of Korean comparative structure sentences. The brain and behavioral response of 15 Korean native speakers were recorded and analyzed during the processing of the comparative sentences with 4 conditions manipulated by the frequency of the predicate (big or small) and the truth-value(true or false). The behavioral results (i.e. judgment time and accuracy) showed that only the effect of the predicate's frequency was significant. However, in the ERP, the predicate's frequency effect and the truth-value effect were reflected in early N400 component(250-400ms) as well as the positivity effect in 400-600ms by the truth-value factor. The immediacy of the truth-value validation and the functional interpretation of N400 related to the truth-value were discussed in overall.

1

내가 찾은 최근 검색어

최근 열람 자료

맞춤 논문

보관함

내 보관함
공유한 보관함

1:1문의

닫기