Hiroshi Noji

Researcher

Artificial Intelligence Research Center
National Institute of Advanced Industrial Science and Technology (AIST)

Contact: hiroshi.noji [at] aist.go.jp

I have moved from NAIST to AIST in April 2018.

Research

My research interest is to uncover structural, or other linguistic biases exist in natural language, and explore the interaction between such insights into language and natural language processing applications.

While end-to-end approaches get much popularity in recent NLP, it seems unlikely that all aspects of language can be learned in a completely data-driven manner. My current interest is to investigate which aspects of language can or cannot be acquired from the data alone, to know how we can make NLP systems more robust, and ultimately humanlike.

I’m also interested in formalisms of syntactic and semantic representations, as well as an efficient algorithm for a particular model, in particular involving structured prediction. Examples of this work include our ACL 2017, EMNLP 2016, and ACL 2015 papers.

My dissertation was about finding a syntactic principle, universal across languages, and applying it to unsupervised grammar induction (unsupervised parsing).

Publications

  1. Learning with Contrastive Examples for Data-to-Text Generation
    *Yui Uehara, *Tatsuya Ishigaki, Kasumi Aoki, Hiroshi Noji, Keiichi Goshima, Ichiro Kobayashi, Hiroya Takamura, and Yusuke Miyao, COLING 2020.
    * Equal contributions.
    [pdf] [bib]

  2. An empirical analysis of existing systems and datasets toward general simple question answering
    Namgi Han, Goran Topic, Hiroshi Noji, Hiroya Takamura, and Yusuke Miyao, COLING 2020.
    [pdf] [bib]

  3. CharacterBERT: Reconciling ELMo and BERT for Word-Level Open-Vocabulary Representations From Characters
    Hicham El Boukkouri, Olivier Ferret, Thomas Lavergne, Hiroshi Noji, Pierre Zweigenbaum, and Jun’ichi Tsujii, COLING 2020.
    [pdf] [bib]

  4. Controlling Contents in Data-to-Document Generation with Human-Designed Topic Labels
    Kasumi Aoki, Akira Miyazawa, Tatsuya Ishigaki, Tatsuya Aoki, Hiroshi Noji, Keiichi Goshima, Ichiro Kobayashi, Hiroya Takamura, and Yusuke Miyao, Computer Speech & Language
    [pdf] [bib]

  5. Learning to Select, Track, and Generate for Data-to-Text
    Hayate Iso, Yui Uehara, Tatsuya Ishigaki, Hiroshi Noji, Eiji Aramaki, Ichiro Kobayashi, Yusuke Miyao, Naoaki Okazaki, and Hiroya Takamura, Journal of Natural Language Processing 2020.
    [pdf] [bib]

  6. An analysis of the utility of explicit negative examples to improve the syntactic abilities of neural language models
    Hiroshi Noji and Hiroya Takamura, ACL 2020.
    [pdf] [code] [bib]

  7. Controlling Contents in Data-to-Document Generation with Human-Designed Topic Labels
    Kasumi Aoki, Akira Miyazawa, Tatsuya Ishigaki, Tatsuya Aoki, Hiroshi Noji, Keiichi Goshima, Ichiro Kobayashi, Hiroya Takamura, and Yusuke Miyao, INLG 2019.
    [pdf] [bib]

  8. Automatic Generation of High Quality CCGbanks for Parser Domain Adaptation
    Masashi Yoshikawa, Hiroshi Noji, Koji Mineshima, and Daisuke Bekki, ACL 2019.
    [pdf] [bib]

  9. Learning to Select, Track, and Generate for Data-to-Text
    Hayate Iso, Yui Uehara, Tatsuya Ishigaki, Hiroshi Noji, Eiji Aramaki, Ichiro Kobayashi, Yusuke Miyao, Naoaki Okazaki, and Hiroya Takamura, ACL 2019.
    [pdf] [bib]

  10. A* CCG Parsing with a Supertag and Dependency Factored Model
    Masashi Yoshikawa, Hiroshi Noji, and Yuji Matsumoto, Journal of Natural Language Processing 2019.
    [pdf] [bib]

  11. Combining Axiom Injection and Knowledge Base Completion for Efficient Natural Language Inference
    Masashi Yoshikawa, Koji Mineshima, Hiroshi Noji, and Daisuke Bekki, AAAI 2019.
    [pdf] [bib]

  12. An Empirical Investigation of Error Types in Vietnamese Parsing
    Quy Nguyen, Yusuke Miyao, Hiroshi Noji, and Nhung Nguyen, COLING 2018.
    [pdf] [bib]

  13. Dynamic Feature Selection with Attention in Incremental Parsing
    Ryosuke Kohita, Hiroshi Noji, and Yuji Matsumoto, COLING 2018.
    [pdf] [bib]

  14. Consistent CCG Parsing over Multiple Sentences for Improved Logical Reasoning
    Masashi Yoshikawa, Koji Mineshima, Hiroshi Noji, and Daisuke Bekki, NAACL-HLT 2018.
    [pdf] [bib]

  15. Can Discourse Relations be Identified Incrementally?
    Frances Yung, Hiroshi Noji, and Yuji Matsumoto, IJCNLP 2017.
    [pdf] [bib] [notes] [dataset]

  16. Effective Online Reordering with Arc-Eager Transitions
    Ryosuke Kohita, Hiroshi Noji, and Yuji Matsumoto, IWPT 2017.
    [pdf] [bib]

  17. Adversarial Training for Cross-Domain Universal Dependency Parsing
    Motoki Sato, Hitoshi Manabe, Hiroshi Noji, and Yuji Matsumoto, CoNLL 2017 (Shared Task). 6-th ranked among 33 participants.

    [pdf] [bib]

  18. A* CCG Parsing with a Supertag and Dependency Factored Model
    Masashi Yoshikawa, Hiroshi Noji, and Yuji Matsumoto, ACL 2017.
    [pdf] [bib]

  19. Multilingual Back-and-Forth Conversion between Content and Function Head for Easy Dependency Parsing
    Ryosuke Kohita, Hiroshi Noji, and Yuji Matsumoto, EACL 2017.
    [pdf] [bib]

  20. Using Left-corner Parsing to Encode Universal Structural Constraints in Grammar Induction
    Hiroshi Noji, Yusuke Miyao, and Mark Johnson, EMNLP 2016.
    [pdf] [slides] [system] [bib]

  21. Jigg: A Framework for an Easy Natural Language Processing Pipeline
    Hiroshi Noji and Yusuke Miyao, ACL 2016 (demo track).
    [pdf] [poster] [system] [bib]

  22. Left-corner Parsing for Dependency Grammar
    Hiroshi Noji and Yusuke Miyao
    Journal of Natural Language Processing 2015. Best paper award of the year.
    [pdf] [bib]

  23. Optimal Shift-reduce Constituent Parsing with Structured Perceptron
    Le Quang Thang, Hiroshi Noji, and Yusuke Miyao, ACL 2015.
    [pdf] [slides] [bib]

  24. Left-corner Transitions on Dependency Parsing
    Hiroshi Noji and Yusuke Miyao, COLING 2014.
    [pdf] [slides] [bib]

  25. Hierarchical Tree-Structured Stick-Breaking Priors
    Hiroshi Noji, Daichi Mochihashi, and Yusuke Miyao,
    NIPS 2013 workshop: Modern Nonparametric Methods in Machine Learning
    [pdf] [poster] [bib]

  26. Improvements to the Bayesian Topic N-gram Models
    Hiroshi Noji, Daichi Mochihashi, and Yusuke Miyao, EMNLP 2013.
    [pdf] [poster] [bib]

Softwares

  • Jigg: A framwork that makes it easy to integrate an NLP tool into existing pipelines.

  • maf: A Python-based management tool for scientific experiments built on waf.

Honors

  • Best Paper Award in The Association for Natural Language Processing 2015 (March 2016)

  • Research Fellowship for Young Scientists, JSPS (Japan Society for the Promotion of Science) (2015)

  • Best Student Award of National Institute of Informatics (March 2015)

  • Incentive Award for Young Researchers of NLP-2014 (March 2014)

Education

  • 2016: PhD at Graduate University for Advanced Studies (SOKENDAI)
    My adviser was Yusuke Miyao, and other thesis committee were: Hiroshi Nakagawa and Daichi Mochihashi (Machine Learning), Edson T. Miyamoto (Psycholinguistics), and Makoto Kanazawa (Formal Linguistics).

  • 2013: M.E at the University of Tokyo (Information Science and Technology)

  • 2011: B.E at Waseda university

Experience

  • April 2018 - : Researcher at Artificial Intelligence Research Center, AIST.

  • April 2016 - March 2018: Assistant Professor at Nara Institute of Science and Technology (NAIST).

  • 2013 - 2016: Research assistant at NII

  • 2014 - 2015: Research fellow at Center for Simulation Sciences in Ochanomizu University.

  • Oct. 2011 - March 2015: Part-time engineer at the Preferred Infrastructure Inc.

  • Summer 2011: Internship at the Preferred Infrastructure Inc.