Organized by the Association for Computer Linguistics special interest group on linguistic data (SIGDAT), Empirical Methods in Natural Language Processing is a prominent conference in the area of Natural Language Processing. The conference on Empirical Methods in Natural Language Processing and 9th International Joint Conference on Natural Language Processing, EMNLP-IJCNLP 2019 was held in Hong Kong from November 3 to 7, 2019. NAVER Corporation attended the conference as a gold-level sponsor.

The following papers were accepted at the EMNLP-IJCNLP 2019.

Subword Language Model for Query Auto-CompletionGyuwan KimSubword language model for query auto-completion (QAC) to decode faster while maintaining close accuracy compared to the character language model baseline, and proposing a new evaluation metric for QAC.arXiv
NL2pSQL: Generating Pseudo-SQL Queries from Under-specified Natural Language QuestionsFuxiang Chen, Seung-won Hwang, Jaegul Choo, Jung-Woo Ha, Sung KimA new dataset which has a better performance in managing and generating tables and is more applicable to the general situation than existing NL2SQLPaper
Mixture Content Selection for Diverse Sequence GenerationJaemin Cho*, Minjoon Seo, Hannaneh HajishirziA new mixture-based model for the generation of diverse queries and summariesarXiv

Please see below for photos!

Minjoon Seo (Clova AI Research) participated in the Machine Reading for Question Answering (MRQA) workshop as a member of the organizing committee. To promote research on MRQA, the workshop sought submissions in two tracks: a research track and a new shared task track. The shared task of the workshop was specifically designed to test how well MRQA systems can generalize to new domains.

Media coverage on NAVER’s achievements at EMNLP-IJCNLP 2019: