Department of Intelligence Science and Technology, Graduate School of Informatics, Kyoto University
Address: S211, Research Bldg. No.9, Yoshida Main Campus, Kyoto University
Yoshida-honmachi, Sakyo-ku, Kyoto, 606-8501, Japan
Mail1: feicheng (at) i, (at) i = @i.kyoto-u.ac.jp
Mail2: kevin.cheng.fei (at) i, (at) i = @gmail.com
Tel: +81 80 3104 4986
I’m a research scientist working on statistical natural language processing. My doctoral research is on using a character-based dependency model to parse the internal structure of Chinese words (LREC2014) and aims to use the internal structure information to improve other Chinese NLP processes. In the following work (ACL2015), we demonstrate how the internal structure information improve state-of-the-art Chinese word segmentation performance.
I'm currently working on extracting temporal information from large-scale news articles for the purpose of generating event timeline automatically. Our work about using a dependency-based neural network for classifying temporal relation is to appear in ACL2017.
Ph.D., Computational Linguistics Apr. 2013 ~ Mar. 2016
Nara Institute of Science and Technology, Nara, Japan
Advisor: Yuji Matsumoto, Kevin Duh
M.E., Computational Linguistics Apr. 2011 ~ Mar. 2013
Nara Institute of Science and Technology, Nara, Japan
Advisor: Yuji Matsumoto
Research Student Oct. 2009 - Sep. 2010
Shizuoka University, Shizuoka, Japan
Advisor: Hiroshi Ishikawa
B.S., Applied Physics Sep. 2001~ Jun. 2005
Donghua University, Shanghai, China
Automatic Error Correction on Japanese Functional Expressions Using Character-based Neural Machine Translation
Liu Jun, Fei Cheng, Yuji Matsumoto
The 32nd Pacific Asia Conference on Language, Information and Computation (PACLIC), oral presentation, Hong Kong, December 2018.
Inducing Temporal Relations from Time Anchor Annotation
Classifying Temporal Relations by Bidirectional LSTM over Dependency Paths
Towards a Consistent Segmentation Level Across Multiple Chinese Word Segmentation Corpora
Synthetic Word Parsing Improves Chinese Word Segmentation
A Hybrid Ranking Approach to Chinese Spelling Check
Parsing Chinese Synthetic Words with a Character-based Dependency Model.
A Hybrid Chinese Spelling Correction Using Language Model and Statistical Machine Translation with Reranking
時刻アノテーションから導出した時間順序関係による時間情報解析
Fei Cheng and Yusuke Miyao
The 24th Annual Meeting of the Association for Natural Language Processing (NLP2018), Okayama, Japan, 2018
経済記事からの不祥事報道検知
Jason Bennett, 野原崇史, Fei Cheng, 石田隆, 宮尾祐介.
The 24th Annual Meeting of the Association for Natural Language Processing (NLP2018), Okayama, Japan, 2018
Program-specific Assistant Professor Apr. 2019 - Present
Department of Intelligence Science and Technology, Graduate School of Informatics, Kyoto University
Involved in constructing Artificial Intelligence in medicine.
Postdoctoral Researcher Sep. 2016 - Mar. 2019
Research Center for Financial Big Data, National Institute of Informatics, Tokyo, Japan
Advisor: Yusuke Miyao
Involved in extracting temporal information from large-scale news articles. Automatic generation of event timeline.
Research Assistant/Researcher Apr. 2012 - Aug. 2016
Computational Linguistics Laboratory, Nara Institute of Science and Technology, Nara, Japan
Advisor: Yuji Matsumoto
Involved in building a Chinese synthetic word dictionary with the internal structure information annotated. Defining the annotation standard, hiring four students to complete the annotation work and investigating the annotation quality, and building a synthetic word parser for analyzing internal structure.
Engineer Jul. 2005 - Sep. 2009
Jiaoda Withub Information Industrial Company Limited, Shanghai, China
JASSO Honors scholarship, japan Apr. 2013 - Mar. 2016
TALLIP 2015~: Reviewer
ACL/COLING 2018: Program Committee Member
NAACL/ACL/EMNLP/AAAI 2019: Program Committee Member