Shun Hasegawa [CV] [LinkedIn] hasegawa.s

English/日本語

---

I'm a third-year Ph.D student at Tokyo Institute of Technology, supervised by Prof. Manabu Okumura.
Previously I recieved my BSc from Tokyo Institute of Technology under the supervision of Prof. Toshiya Itoh.
My research interest is machine learning for natural language processing, and particularly Natural Language Understanding.

Publications

Journal

• Optimal Online Algorithms for the Multi-Objective Time Series Search Problem [paper]
  Shun Hasegawa and Toshiya Itoh.
  Journal of Theoretical Computer Science, Springer, Vol. 718, pp. 58-66, 2018.

International Conferences

• Japanese Sentence Compression with a Large Training Dataset [paper]
  Shun Hasegawa, Yuta Kikuchi, Hiroya Takamura and Manabu Okumura.
  In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (ACL). 2017.
• Optimal Online Algorithms for the Multi-Objective Time Series Search Problem [paper]
  Shun Hasegawa and Toshiya Itoh.
  In Proceedings of The 10th International Conference and Workshops on Algorithms and Computing (WALCOM). 2016.

Domestic Conferences

• Extractiveness-based dataselection for abstractive sentence summarization [paper]
  Shun Hasegawa, Hidetaka Kamigaito, Manabu Okumura.
  In IPSJ Natural Language Processing (2019.8), Young Researcher Award
• +6 papers (in Japanese)
共参照情報を活用した映画のあらすじ要約生成
川原田 将之, 長谷川 駿, 上垣外 英剛, 高村 大也, 奥村 学.
言語処理学会 第27回年次大会 (2021.3)
傾聴の応答で繰り返される語句の検出向上
川本 稔己, 長谷川 駿, 上垣外 英剛, 船越 孝太郎, 奥村 学.
言語処理学会 第27回年次大会 (2021.3)
事前学習と汎化タグによる方言翻訳の精度向上 [paper]
長谷川 駿, 田中 駿, 山本 悠二, 高村 大也, 奥村 学.
情報処理学会 第233回自然言語処理研究発表会 (2017.10)
文圧縮を活用したヘッドライン生成 [paper]
長谷川 駿, 平尾 努, 奥村 学, 永田 昌明.
言語処理学会 第23回年次大会 (2017.3)
大規模データを用いた日本語文圧縮 [paper]
長谷川 駿, 菊池 悠太, 高村 大也, 奥村 学.
言語処理学会 第23回年次大会 (2017.3)
Best Possible Online Algorithms for the Multi-Objective Time Series Search Problem [paper]
Shun Hasegawa and Toshiya Itoh.
LAシンポジウム2016 (2016.7)

Education

2018.04 - present     Ph.D in Engineering
Tokyo Institute of Technology
  Supervisor: Manabu Okumura
  Theme: Improvement of single document summarizer


2016.04 - 2018.03    M.S. in Engineering
Tokyo Institute of Technology
  Supervisor: Manabu Okumura
  Theme: Japanese Sentence Compression with a Large Training Dataset


2012.04 - 2016.03    B.S. in Computer Science
Tokyo Institute of Technology
  Supervisor: Toshiya Itoh
  Theme: Optimal Online Algorithms for the Multi-Objective Time Series Search Problem

Experience

2020.01 - present     Research Assistant
Tokyo Institute of Technology, Okumura-Takamura laboratory
  Reduced training time to 50% without performance regression by the proposed data selectionmethod in automatic text summarization
  Implemented a novel summarization neural model from scratch (Pytorch)


2016.08 - 2016.12    Research Intern
NTT Communication Science Laboratories
  Supervisor: Tsutomu Hirao
  Adapted sentence compression for pre-processing of sentence summarization


2016.04 - 2018.03    Admin Group Leader
Tokyo Institute of Technology, Laboratory Server Admin
  Formulated server specifications that suit the budget and server usage (gpu, cpu, file servers)
  Set up and maintained 30 servers (gpu, cpu, file, web, gateway, dns, nis and dhcp servers)
  Reorganized admin group to minimize mental and time burden



2016.01 - 2019.02    Research Assistant
Tokyo Institute of Technology, Okumura-Takamura laboratory
  Adapted the rule-based method using syntactic information from English to Japanese
  Developed a neural dialect translation method using low-resource data (1k parallel sentences)


2014.04 - 2016.03    Software Engineer
Shakehands Co., Ltd.
  Improved CVR by modifying wordpress-driven web pages and landing page
  Analyzed web pages using Google Analytics
  Work experience in a venture of several employees

Skills

• Programming: Python3, Bash, Pytorch, Chainer, Scikit-learn, numpy, C, Java, php, javascript, html, css, ganglia, nis, autofs, zfs, dhcpd, iptables, appach, quagga, docker
  
• Languages: English (Proficient), Japanese (Native), Mandarin Chinese (Elementary)
  
• Research Theme: Automatic Text Summarization (Abstractive Sentence Summarization, Headline Generation, Sentence Compression, Abstractive Single Document Summarization, Sentence Selection), Automatic Dialect Translation, Grammatical Error Correction
  
• ML, NLP: Deep Learning (LSTM, Transformer, Bert), SVM, SMT, word2vec, ILP, Sequence Labeling, Natural Language Understanding, Text Generation, Data Filtering, Data Selection, Syntactic Analysis, Discourse Structure Analysis, Tf-idf, etc
  
• Hobby: Tennis, Baseball, Table tennis, Fishing, Billiards, Board game, Travel, etc