Yasuhisa Yoshida

Profile

• Interested in a theory of natural language processing and machine learning, their applications and making softwares for them.
• Problem solving skills including problem definition and decision making.
• Computer and Internet skills.
• Excellent organizational, coordinating skills and ability to motivate others.

Research Interests

• Natural Language Processing
• Sentiment Analysis
• Automatic Summarization
• Discourse Structure, Discourse Parsing
• Topic Model
• Word Segmentation
• Machine Learning
• Probabilsitic Generative Model
• Nonparametric Bayes
• Transfer Learning
• Nonparametric Bayes
• Unsupervised Learning, Semi-Supervised Learning
• Variational Bayes, MCMC

Education

• MS Graduate School of Information Science, Nara Institute of Science and Technology 2010 April - Current
• Scholarship of Graduate School of Information Science, Nara Institute of Science and Technology
• MS Social Systems Engineering, University of Tsukuba April 2009 - March 2010
• BS Management Science and Engineering, University of Tsukuba April 2005 - March 2009
• Thesis title : Nonparametric multiple regression under Nadaraya-Watson estimator
• Thesis advisor : Associate Professor, Yuichiro Kanazawa

Work Experience

Research Assistant

• Database Center for Life Science( http://dbcls.rois.ac.jp/en/ )
• 4/2009 - 3/2010
• Developing a system of automatically recognizing domain specific terms from corpus.
• Main objective of this is improve an accuracy of full text search for protein nucleic acid and enzyme( http://lifesciencedb.jp/pne/ ).
• Another application using dictionary of domain specific terms is "Trending words in each year for protein nucleic acid and enzyme"( http://g86.dbcls.jp/~syou6162/pne/ , in Japanese).

Summer Internship

• Microsoft Research Asia
• 2011/9 - 2011/11
• Under guidance of Chin-Yew Lin
• NTT Communication Science Laboratories, Japan( http://www.kecl.ntt.co.jp/rps/index.html )
• 8/2010 - 9/2010, 11/2010 - 3/2011
• Summer Research Intern: Transfer Learning for Natural Language Processing
• Under guidance of Tsutomu Hirao
• Hatena Co., Ltd( http://www.hatena.ne.jp/company/ , in Japanese)
• 8/2009
• Developed a spam filter for the comment and trackback of blogs.
• Outstanding performance award in the company.
• Data4C's Co., Ltd( http://www.data4cs.co.jp/ , in Japanese)
• 7/2007 - 8/2007
• Worked as a data mining consultant for the questionnaires from web.

Publications

• Yasuhisa Yoshida, Tsutomu Hirao, Tomoharu Iwata, Masaaki Nagata, Yuji Matsumoto, Transfer Learning for Multiple-Domain Sentiment Analysis: Identifying Domain Dependent/Independent Word Polarity, In Proceedings of the Twenty-Fifth AAAI Conference on Artificial Intelligence, San Francisco, USA, August 2011. ( paper, presentation, poster )

Awards

• Outstanding Paper Award, the 17th Annual Meeting of the Association for Natural Language Processing, Japan. (2011.3)

Computer Skills

• Languages : C++, Gnu R, Ruby, Perl, Python
• Platforms : Mac OS X, Linux(Ubuntu, Vine Linux), Windows XP

Software

• nw
• Nonparametric regression with Nadaraya-Watson estimator
• This package enables you to do a regression with Nadaraya-Watson estimator especially in two dimensions
• oll-ruby
• This is a Ruby binding of oll , a library specialized for large-scale, but sparse, learning tasks such as Natural Language Processing
• local_hatena_search
• Simple blog system supporting hatena format .
• This uses SuffixArray to help searching quickly.
• Actually, this system is working here .
• realtime_search_engine
• Simple realtime search for twitter's timeline.
• SimpleHMM
  • Simple implementation of Hidden Markov Model in C++.
• fobos
  • Implementation of Support Vector Machine with L1 regularization and Logistic Regression with L1 regularization optimized by FOBOS(Forward-Backward Splitting)

Activities

Tsukuba.R

• 7/2007 - current
• User group of Gnu R organized by myself
• Gave a lot of technical presentation about Gnu R.
• Gave a presentation about Tsukuba.R to Japanese Gnu R user group held at the institute of statistical mathematics( http://www.ism.ac.jp/index_e.html ).
• Introduced in Local R User Group Directory by Revolutions( http://blog.revolution-computing.com/local-r-groups.html ).

RubyKaigi

• 7/2008 and 7/2009
• RubyKaigi is a international conference for Ruby.
• Worked as a KaigiFreaks(volunteer staff for these conferences).

Other Achivements

• Planned and lead a machine learning study group, which was held weekly to share and discuss various reseach topics of the participants(one year).