2019/08/25 by Kang Min Yoo, Yoo, Kang Min, Taeuk Kim +3
Computer Science · #Computation and Language (cs.CL) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Natural Language Processing Techniques #Speech Recognition and Synthesis #Topic Modeling
paper · pdf · doi:10.48550/arxiv.1908.09282
openalex publication_date 2019/08/25 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We propose a simple yet effective approach for improving Korean word\nrepresentations using additional linguistic annotation (i.e. Hanja). We employ\ncross-lingual transfer learning in training word representations by leveraging\nthe fact that Hanja is closely related to Chinese. We evaluate the intrinsic\nquality of representations learned through our approach using the word analogy\nand similarity tests. In addition, we demonstrate their effectiveness on\nseveral downstream tasks, including a novel Korean news headline generation\ntask.\n