Kenji Imamura (今村 賢治)


A researcher of natural language processing.
My research interests include: 1) machine translation, 2) dependency parsing, 3) predicate-argument structure analysis, and 4) grammatical error correction in Japanese.

Recent Publications


2019

Usage of Long Context in Dialogue Translation

  • in Japanese. 「対話翻訳における長距離文脈の利用」
  • Kenji Imamura and Eiichiro Sumita.
  • The 25th Annual Meeting of Assoc. for Natural Language Processing, pp.550-553, March 2019.
  • [paper (J-stage) / bib]

2018

Usage of Multiple Models in Neural Machine Translation

  • in Japanese. 「ニューラル機械翻訳における複数モデルの利用」
  • Kenji Imamura and Eiichiro Sumita.
  • Japio YEAR BOOK 2018, pp.304-308, Nov. 2018.
  • [paper (Japio Web Page) / bib]

Enhancement of Encoder and Attention Using Target Monolingual Corpora in Neural Machine Translation

  • Kenji Imamura, Atsushi Fujita, and Eiichiro Sumita.
  • The 2nd Workshop on Neural Machine Translation and Generation (NMT-2018) in ACL-2018, pp.55-63, Jul. 2018.
  • [paper (ACL Anthology) / bib]

NICT Self-Training Approach to Neural Machine Translation at NMT-2018

  • Kenji Imamura and Eiichiro Sumita.
  • The 2nd Workshop on Neural Machine Translation and Generation (NMT-2018) in ACL-2018, pp.110-115, Jul. 2018.
  • [paper (ACL Anthology) / bib]

Multilingual Parallel Corpus for Global Communication Plan

  • Kenji Imamura and Eiichiro Sumita.
  • 11th edition of the Language Resources and Evaluation Conference (LREC-2018), pp.3453-3458, May 2018.
  • [paper / bib]

Training Strategy for Encoder-Decoder Using Monolingual Corpura and Back-Translation

  • in Japanese. 「単言語コーパスと逆翻訳を用いたエンコーダー・デコーダーの訓練法」
  • Kenji Imamura, Atsushi Fujita, and Eiichiro Sumita.
  • The 24th Annual Meeting of Assoc. for Natural Language Processing, pp.29-32, March 2018.
  • [paper (J-stage) / bib]

Multilingual Parallel Corpus for Global Communication Plan

  • in Japanese. 「グローバルコミュニケーション計画のための多言語パラレルコーパス」
  • Kenji Imamura and Eiichiro Sumita.
  • The 24th Annual Meeting of Assoc. for Natural Language Processing, pp.512-515, March 2018.
  • [paper (J-stage) / bib]

2017

Ensemble and Reranking: Using Multiple Models in the NICT-2 Neural Machine Translation System at WAT2017

  • Kenji Imamura and Eiichiro Sumita.
  • 4th Workshop on Asian Translation (WAT2017) in IJCNLP-2017, pp.127-134, Nov. 2017.
  • [paper (ACL Anthology) / bib]

Usage of Multiple Models by Ensemble and Bi-directional Reranking in Neural Machine Translation

  • in Japanese. 「双方向リランキングとアンサンブルを併用したニューラル機械翻訳における複数モデルの利用法」
  • Kenji Imamura and Eiichiro Sumita.
  • IPSJ SIG Technical Report 2017-NL-233, No.9, Oct. 2017.
  • [paper (IPSJ digital library) / bib]

Multi-domain Adaptation for Statistical Machine Translation Based on Feature Augmentation

  • in Japanese. 「素性空間拡張法に基づくフレーズベース統計翻訳のマルチドメイン適応」
  • Kenji Imamura and Eiichiro Sumita.
  • Journal of Natural Language Processing, Vol.24, No.4, pp. 597-618, Sep. 2017.
  • [paper (J-STAGE) / bib]

History of Machine Translation

  • in Japanese. 「機械翻訳の歴史」
  • Kenji Imamura.
  • Encyclopedia of Artificial Intelligence, Sec.8-29, pp.668-671, Kyoritsu Publisher, ISBN:978-4-320-12420-2, Jul. 2017.
  • [Kyoritsu Publisher / bib]

Introducing Monolingual Corpora to Neural Machine Translation via Pseudo-Source Genaration

  • in Japanese. 「疑似原文生成によるニューラル機械翻訳への単言語コーパスの導入」
  • Kenji Imamura and Eiichiro Sumita.
  • The 23rd Annual Meeting of Assoc. for Natural Language Processing, pp.1023-1026, March 2017.
  • [paper / bib]

2016

NICT-2 Translation Sysytem for WAT2016: Applying Domain Adaptaion to Phrase-based Statistical Machine Translation

  • Kenji Imamura and Eiichiro Sumita.
  • The 3rd Workshop on Asian Translation (WAT2016) in COLING 2016, pp. 126-132, December 2016.
  • [paper (ACL Anthology) / bib]

Multi-domain Adaptation for Statistical Machine Translation Based on Feature Augmentation

  • Kenji Imamura and Eiichiro Sumita.
  • The Association for Machine Translation in the Americas (AMTA 2016), pp. 79-92, October 2016.
  • [paper / bib]

Domain Adaptation Based on Corpus Concatenation and Feature Augmentation for Statistical Machine Translation

  • in Japanese. 「コーパス結合モデルと素性空間拡張に基づく統計翻訳のドメイン適応」
  • Kenji Imamura and Eiichiro Sumita.
  • Japio YEAR BOOK 2016, pp.252-259, November 2016.
  • [paper (Japio WEB page) / bib]

Hierarchical Phrase Alignment Harmonized with Parsing

  • Kenji Imamura.
  • Francis Bond, Timothy Baldwin, Kentaro Inui, Shun Ishizaki, Hiroshi Nakagawa, and Akira Shimazu editors.
  • Readings in Japanese Natural Language Processing, Chapter 11, CSLI Publications, ISBN:978-1-575-86753-3, June 2016.
  • [CSLI page / bib]

Multi-domain Adaptation for Statistical Machine Translation Using Feature Augmentation and Corpus Concatenation

  • in Japanese. 「素性空間拡張とコーパス結合を併用した統計翻訳のマルチドメイン適応」
  • Kenji Imamura and Eiichiro Sumita.
  • The 22nd Annual Meeting of Assoc. for Natural Language Processing, pp.203-206, March 2016.
  • [paper / bib]

2015

Optimization Strategies for Better Fitting Training Data in Statistical Machine Translation

  • in Japanese. 「訓練データにより適合した統計翻訳最適化戦略」
  • Kenji Imamura and Eiichiro Sumita.
  • IPSJ SIG Technical Report 2015-NL-222, No.3, July 2015.
  • [paper (IPSJ digital library) / bib]

History of Machine Translation (that I've never considered)

  • in Japanese.「(考えてもみなかった)機械翻訳の歴史」
  • originally presented at the 12th Kansai MT Study Session (July 4th, 2015).
  • [slides(pdf) / slides(pptx)]

Predicate-Argument Structure Analysis and Zero-Anaphora Resolution for Dialogue Analysis

  • in Japanese. 「対話解析のためのゼロ代名詞照応解析付き述語項構造解析」
  • Kenji Imamura, Ryuichiro Higashinaka, and Tomoko Izumi.
  • Journal of Natural Language Processing, Vol.22, No.1, pp.3-26, March 2015.
  • [paper (J-STAGE) / bib]