Ayana Niwa

Research Scientist at Megagon Labs Tokyo
Part-time Researcher at Tokyo Institute of Technology

Research Interest: Natural Language Generation (NLG)

My CV can be found here (last updated: Mar. 15, 2024).

Work Experience

2023/04 - present
Megagon Labs
Research scientist
2023/06 - present
2021/04 - 2023/03
2019/06 - 2023/03
Megagon Labs
Research Internship
2018/10 - 2023/03
CyberAgent, Inc.
Collaborative project
2018/10 - 2023/03
Tokyo Institute of Technology
Research Assistant

Publication

Journal (Refereed)
  • Ayana Niwa, Sho Takase, Naoaki Okazaki. Nearest Neighbor Non-autoregressive Text Generation. Journal of Information Processing, 2023. URL
  • Ayana Niwa, Naoaki Okazaki, Kohei Wakimoto, Keisuke Nishiguchi, Masataka Mouri. Construction of a Corpus of Rhetorical Devices in Slogans and Structural Analysis of Antitheses. ACM Transactions on Asian Low-Resource Language Information Processing (TALLIP), 2021. URL
International Conference (Refereed)
  • Masahiro Kaneko, Sho Takase, Ayana Niwa, Naoaki Okazaki. Interpretability for Language Learners Using Example-Based Grammatical Error Correction. Proceedings of the 60th Annual Conference of the Association for Computational Linguistics (ACL), 2022. URL
  • Ayana Niwa, Keisuke Nishiguchi, Naoaki Okazaki. Predicting Antonyms in Context using BERT. Proceedings of the 14th International Conference on Natural Language Generation (INLG2021), , 2021. URL
Preprint
  • Ayana Niwa, Hayate Iso. AmbigNLG: Addressing Task Ambiguity in Instruction for NLG. arxiv preprint, 2024. Paper
Article
  • Ayana Niwa, Sho Yokoi, Junya Takayama, Itsumi Saito. NLP Colloquium. Journal of Natural Language Processing, 2024. Paper
Domestic Conference and Symposium (Unrefereed)
  • Ayana Niwa, Hayate Iso. Addressing Ambiguity in Instruction for NLG. The 30th Annual Meeting of the Association of Natural Language Processing (NLP2024), Hyogo, Japan. (in Japanese) Paper
  • Ayana Niwa, Naoaki Okazaki. An Investigation of Effectiveness of the Nearest Neighbor Distribution in T5 Model. The 29th Annual Meeting of the Association for Natural Language Processing (NLP2023), Okinawa, March 2023. (in Japanese) Paper
  • Daisuke Taniguchi, Kohei Wakimoto, Ayana Niwa, Naoaki Okazaki. An Investigation of the Directional Dependency of Text Generation in Large Language Models. The 29th Annual Meeting of the Association for Natural Language Processing (NLP2023), Okinawa, March 2023. (in Japanese) Paper
  • Daisuke Taniguchi, Kohei Wakimoto, Kazuya Kuroda, Shunyo Kawamoto, Keisuke Nishiguchi, Ayana Niwa, Naoaki Okazaki. Advertisement text creation support using product reviews and features. The 36th Annual Conference of the Japanese Society for Artificial Intelligence (JSAI2023), Online, June 2023. (in Japanese) Paper
  • Ayana Niwa, Sho Takase, Naoaki Okazaki. Non-autoregressive Generation using the Nearest Neighbor. The 28th Annual Meeting of the Association for Natural Language Processing (NLP2022), online, March 2022. (in Japanese) Paper
  • Ryogo Ishikawa, Ayana Niwa, Mizuki Sakae, Naoaki Okazaki. Robust Dependency Parsing for the Omission of a Post-positional Particle using Pseudo Training Data. The 28th Annual Meeting of the Association for Natural Language Processing (NLP2022), online, March 2022. (in Japanese) Paper
  • Ayana Niwa, Hiroshi Matsuda. A Study for Sentiment Analysis Accepting the Diverse of Emotion Sensitivities. The 35th Annual Conference of the Japanese Society for Artificial Intelligence (JSAI2021), online, June 2021. (in Japanese) Paper
  • Ayana Niwa, Keisuke Nishiguchi, Naoaki Okazaki. Antonym Filling Task in Context using BERT. The 27th Annual Meeting of the Association for Natural Language Processing (NLP2021), online, March 2021. (in Japanese) Paper
  • Natsumi Nobori, Tatsuya Hiraoka, Ayana Niwa, Keisuke Nishiguchi, Naoaki Okazaki. Advertisement Slogan Generation Considering Company Information. The 27th Annual Meeting of the Association for Natural Language Processing (NLP2021), online, March 2021. (in Japanese) Paper
  • Ayana Niwa, Keisuke Nishiguchi, Naoaki Okazaki. Advertising Slogan Generation with Antitheses by using Masked Language Models. The 15th Symposium of Young Researcher Association for NLP Studies (YANS2020), online, September 2020. (in Japanese)
  • Natsumi Nobori, Tatsuya Hiraoka, Ayana Niwa, Keisuke Nishiguchi, Naoaki Okazaki. Advertisement Slogan Generation Considering Company Information. The 15th Symposium of Young Researcher Association for NLP Studies (YANS2020), online, September 2020. (in Japanese)
  • Ayana Niwa, Kohei Wakimoto, Keisuke Nishiguchi, Masataka Mouri, Naoaki Okazaki. Structural Analysis of Antitheses in Slogan using Span Candidate Refinement with Word Correspondence. The 34th Annual Conference of the Japanese Society for Artificial Intelligence (JSAI2020), Kumamoto online, June 2020. (in Japanese) URL
  • Ayana Niwa, Kohei Wakimoto, Keisuke Nishiguchi, Masataka Mouri, Naoaki Okazaki. Structural Analysis of Antitheses in Advertising Slogans. The 26th Annual Meeting of the Association for Natural Language Processing (NLP2020), Ibaraki online, March 2020. (in Japanese) Paper
  • Ayana Niwa, Naoaki Okazaki, Keisuke Nishiguchi, Chihiro Kameyama, Masataka Mouri. A Study Toward Automatic Advertising Slogan Generation Considering Rhetorical Devices. The 14th Symposium of Young Researcher Association for NLP Studies (YANS2019), Hokkaido, August 2019. (in Japanese) Poster
  • Ayana Niwa, Naoaki Okazaki, Keisuke Nishiguchi, Chihiro Kameyama, Masataka Mouri. Analysis Toward Automatic Generation of Advertising Slogans. The 25th Annual Meeting of the Association for Natural Language Processing (NLP2019), Aichi, March 2019. (in Japanese) Paper

Education

2020/04 - 2023/03
Ph.D. in Engineering, Tokyo Institute of Technology
Adviser: Prof. Naoaki Okazaki
2018/04 - 2020/03
Master of Engineering, Tokyo Institute of Technology
Adviser: Prof. Naoaki Okazaki
2014/04 - 2018/03
Bachelor of Science and Technology, Sophia University, Japan
Adviser: Prof. Takashi Irohara

Talk

Others

Scholarship and Grants-in-Aid
  • Research Fellowship for Young Scientists by Japan Society for the Promotion of Science (JSPS)
    4,800,000 JPY / APPROX 45,000 USD per year (2021/04 - 2023/03)
    The most prestigious fellowship for Ph.D. students in Japan [Acceptance Rate < 20%]
  • Research Grant for JSPS Fellows (Grant Number: 21J13602)
    1,500,000 JPY / APPROX 13,500 USD (2021/04 - 2023/03)
  • NeurIPS 2019 Participation Support by The Japanese Society for Artificial Intelligence
    Full support of travel expense and registration fees (2019/12)
  • Tokyo Tech Tsubame Scholarship for Doctoral Students
    480,000 JPY / APPROX 3,200 USD per year (2020/04 - 2021/03)
Awards
  • Young Researcher’s Encouragement Award, NLP2024 (18/427 papers)
  • Committee Special Award, NLP2022 (16/365 papers)
  • Nomination Award, 2021 Microsoft Research Asia Fellowship