형태소 기반 효율적인 한국어 단어 임베딩

다국어 초록 (Multilingual Abstract)

Previous word embedding models such as word2vec and Glove are not able to learn the internal structure of words. This is a serious limitation for agglutinative languages with morphology such as Korean. In this paper, we propose a new model which is an expansion of the previous skip-gram model. This defines each word vector as a sum of its morpheme vectors and hence, learns the vectors of morphemes. To test the efficiency of our embedding, we conducted a word similarity test and a word analogy test. Furthermore, using our trained vectors on other NLP tasks, we tested how much performance actually had been enhanced.

참고문헌 (Reference)

1 "WordSim353"

2 "Twitter-korean-text"

3 "Tensorflow"

4 Sanghyuk Choi, "On Word Embedding Models and Parameters Optimized for Korean" (15) : 2016

5 "Naver sentiment movie corpus"

6 Cui, Qing, "Knet: A general framework for learning word embedding using morphological knowledge" 34 (34): 2015

7 Levy, Omer, "Improving distributional similarity with lessons learned from word embeddings" 3 : 211-225, 2015

8 Pennington, Jeffrey, "Glove: Global Vectors for Word Representation" 14 : 1532-1543, 2014

9 Schnabel, Tobias, "Evaluation methods for unsupervised word embeddings" 298-307, 2015

10 Bojanowski, Piotr, "Enriching word vectors with subword information"

11 Mikolov, Tomas, "Efficient estimation of word representations in vector space"

12 Baroni, Marco, "Don't count, predict! A systematic comparison of context-counting vs. context-predicting semantic vectors" 1 : 238-247, 2014

13 Mikolov, Tomas, "Distributed representations of words and phrases and their compositionality" 3111-3119, 2013

14 Kim, Yoon, "Convolutional neural networks for sentence classification"

15 Lazaridou, Angeliki, "Compositionally Derived Representations of Morphologically Complex Words in Distributional Semantics" (1) : 1517-1526, 2013

16 양희정, "A Study on Word Vector Models for Representing Korean Semantic Information" 한국음성학회 7 (7): 41-47, 2015