Tatsuya Aoki
[CV]
[github]
[LinkedIn]
tatsuya_aoki
I'm currently working at Apple Inc. as an input method engineer.
Also, I'm a Ph.D candidate at
Tokyo Institute of Technology, supervised by
Prof. Hiroya Takamura and
Prof. Manabu Okumura.
Previously I recieved my BSc from
The University of Tsukuba (Japanese web page) under the supervision of
Prof. Takashi Inui.
My research interest is in machine learning for natural language processing, and I am particularly interested in making machine more robust to lexical diversity as in social media.
Outside the university, I have worked on social media analysis and natural language generation from numeric data (data2text).
Publications
Journals and others
-
Detecting Non-standard Word Usages from Social Media. (in Japanese)
Tatsuya Aoki, Ryohei Sasano, Hiroya Takamura and Manabu Okumura.
Journal of Natural Language Processing, The Association for Natural Language Processing, Vol. 26, No. 2, 2019.
-
Detecting Earthquake Survivors with Serious Mental Affliction.
Tatsuya Aoki, Katsumasa Yoshikawa, Tetsuya Nasukawa, Hiroya Takamura and Manabu Okumura.
Computational Linguistics, Springer Nature, Vol. 781, pp. 3-14, 2018.
International Conferences
-
Neural Text Normalization Leveraging Similarities of Strings and Sounds
Riku Kawamura, Tatsuya Aoki, Hidetaka Kamigaito, Hiroya Takamura and Manabu Okumura.
In Proceedings of the Conference on Computational Linguistics (COLING). 2020.
-
Controlling Contents in Data-to-Document Generation with Human-Designed Topic Labels [paper]
Kasumi Aoki, Akira Miyazawa, Tatsuya Ishigaki, Tatsuya Aoki, Hiroshi Noji, Keiichi Goshima, Ichiro Kobayashi, Hiroya Takamura and Yusuke Miyao.
In Proceedings of the Conference on Natural Language Generation (INLG). 2019.
-
Generating Market Comments Referring to External Resources [paper]
Tatsuya Aoki, Akira Miyazawa, Kasumi Aoki, Keiichi Goshima, Tatsuya Ishigaki, Ichiro Kobayashi, HiroyaTakamura and Yusuke Miyao.
In Proceedings of the Conference on Natural Language Generation (INLG). 2018.
-
Distinguishing Japanese Non-standard Usages from Standard Ones [paper]
Tatsuya Aoki, Ryohei Sasano, Hiroya Takamura and Manabu Okumura.
In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP). 2017.
-
Detecting Earthquake Survivors with Serious Mental Affliction.
Tatsuya Aoki, Katsumasa Yoshikawa, Tetsuya Nasukawa, Hiroya Takamura and Manabu Okumura.
In Proceedings of the Conference of the Pacific Association for Computational Linguistics (PACLING). 2017.
Domestic Conferences
+
-
時系列数値データにおける変化要因の記述
Tatsuya Aoki, Akira Miyazawa, Kasumi Aoki, Keiichi Goshima, Ichiro Kobayashi, Hiroya Takamura and Yusuke Miyao.
言語処理学会 第24回年次大会 (2018.3)
-
ソーシャルメディアにおける俗語の検出 [paper]
Tatsuya Aoki, Ryohei Sasano, Hiroya Takamura and Manabu Okumura.
言語処理学会 第23回年次大会 (2017.3)
-
震災被災者のソーシャルメディア分析を通じたメンタルケアの可能性
Tatsuya Aoki, Tetsuya Nasukawa, Katsumasa Yoshikawa, Hiroya Takamura and Manabu Okumura.
第10回テキストマイニングシンポジウム (2017.2)
Experience
- 2020.10 - Present Input Method Engineer
Apple Inc.
- 2019.06 - 2020.07 Research Intern
- 2019.07 - 2019.11 ML/NLP Intern
Apple Inc.
- 2018.06 - 2019.05 Visiting Student
- 2017.05 - 2018.05 Technical Staff
- 2016.12 - 2017.02 Research Assistant
- 2016.09 - 2016.11 Research Intern
Miscellaneous
PyTorch implementation
-
Dependency Parsing as Head Selecton, Zhang et al., EACL 2017 [github]
with an emperical result that ELMo embedding can gain accuracy.
-
context2vec: Learning Generic Context Embedding with Bidirectional LSTM, Melamud et al., CoNLL 2016 [github]
with 3x speed up compared to the original implementation.