[
日本語 | English ]
Jin SAKUMA
Profile
Affiliation:
The University of Tokyo, Graduate School of Information Science and Techonology Yoshinaga Lab
Educational Background:
2017/4 - present: The University of Tokyo, Graduate School of Information Science and Techonology
2014/9 - 2016/12: The University of Michigan - Flint, Mathematics
2012/4 - 2014/6: Milford High School
2011/4 - 2012/3: Koryo High School, Hokuto City, Yamanashi, Japan
Area of Study
- Natural Language Processing
- Machine Learning
Publication
Refereed Publications
Multilingual model using cross-task embedding projection
Jin Sakuma, Naoki Yoshinaga
The SIGNLL Conference on Computational Natural Language Learning (CoNLL 2019)
(to appear)
Multilingual model using cross-task embedding projection
Jin Sakuma, Naoki Yoshinaga
57th Annual Meeting of the Association for Computational Linguistics, Student Research Workshop (ACL SRW 2019)
(non-archival option is used for future publication)
Unsupervised Cross-lingual Word Embeddings Based on Subword Alignment
Jin Sakuma, Naoki Yoshinaga
20th International Conference on Computational Linguistics and Intelligent Text Processing (CICLing 2019)
A Bag of Useful Tricks for Practical Neural Machine Translation: Embedding Layer Initialization and Large Batch Size
Masato Neishi, Jin Sakuma, Satoshi Tohda, Shonosuke Ishiwatari, Naoki Yoshinaga, Masashi Toyoda
2018 Asian Universities Alliance Postgraduate Academic Forum (AUAPAF 2018)
Other Publications
単語分散表現のタスク横断写像に基づく高精度多言語モデル
(Multilingual models based on cross-task projection of word embeddings)
佐久間 仁, 吉永 直樹
言語処理学会 第25回年次大会(NLP 2019)
若手奨励賞受賞
タスクに特化した多言語単語分散表現を用いた多言語モデル
(Multilingual models using task-specific cross-lingual word embeddings)
佐久間 仁, 吉永 直樹
NLP若手の会 第13回シンポジウム(YANS 2018)
表層類似性を用いた多言語単語分散表現の教師なし学習手法
(Unsupervised Learning of Multilingual Word Embedding via Surface Similarity)
Jin SAKUMA, Naoki Yoshinaga (IIS)
The 233th Meeting on Natural Language Processing, IPSJ
ニューラル機械翻訳における埋め込み層の教師なし事前学習
(Unsupervised Embedding Layer Initialization for Neural Machine Translation)
Masato Neishi, Jin Sakuma, Satoshi Tohda, Shonosuke Ishiwatari, Naoki Yoshinaga, Masashi Toyoda
The 233th Meeting on Natural Language Processing, IPSJ
A Bag of Useful Tricks for Practical Neural Machine Translation: Embedding Layer Initialization and Large Batch Size
Masato Neishi, Jin Sakuma, Satoshi Tohda, Shonosuke Ishiwatari, Naoki Yoshinaga, Masashi Toyoda
The 4th Workshop on Asian Translation
Work Experience
- Software Enigineer Internship
- C.A. Mobile
- 2016/8
- Supplimental Instruction Leader
- The University of Michigan
- 2016/9 - 2016/12
- Research Internship
- Studio Ousia
- 2017/6 -