Japanese Culture and AI Symposium 2019
AI for Reading Kuzushiji is Now Ready!

In Japan, historical materials such as pre-modern books, old documents, and historical records have been preserved for more than a thousand years, and the amount of historical materials are estimated to be hundreds of millions, which is rare in the world. Most modern Japanese people, however, are unable to read historical materials written in kuzushiji, and reading and understanding a large amount of kuzushiji became an urgently important issue. This symposium introduces the forefront of kuzushiji research spreading around the world, discussing past and present to the future of research that uses AI (artificial intelligence) to read and understand kuzushiji.

Kaggle competition: Kuzushiji Recognition

Basic Information

Date November 11 (Mon) 2019, 10:00-17:45 (open at 9:30)
Venue Hitotsubashi Hall
2-1-2, Hitotsubashi, Chiyoda-ku, Tokyo
Google Maps
Fee Free
Registration Registration is required: Registration page
Please register early, because registration may be closed when the number exceeds the capacity of the venue.
Language Session 1-2 and keynote: Japanese only
Session 3-4: Simultaneous translation to English
Live Broadcast / Video Archives This symposibum will offer the live broadcast and video archives. Please see the following YouTube Channel if you cannot join the symposium or would like to see it later.
ROIS-DS CODH YouTube Channel

News

2019-11-08

We updated information about the live broadcast and video archives on YouTube.

2019-10-07

The number of registration exceeded 200.

2019-10-06

We prepared the flyer of the symposium. Please feel free to download and use it.

2019-09-13

Registration has started.

Program

(Note) Talks with the Japanese title will not have English translation. Demonstrations may be presented in English.

9:30 Open the venue
10:00-10:10 Opening remarks Asanobu Kitamoto (ROIS-DS Center for Open Data in the Humanities / National Institute of Informatics)
Session 1 - Japanese literate culture and AI
10:10-10:35 木簡情報のオープンデータ化と文字画像DB連携の強化 Hajime Baba (Nara National Research Institute for Cultural Properties)
10:35-11:00 東京大学史料編纂所における字形データの蓄積経緯と花押データへの展開 Satoshi Inoue (Historiographical Institute, the University of Tokyo)
11:00-11:25 新たな検索機能提供のための調査研究活動―次世代デジタルライブラリーを中心とした近年の取組紹介― Toru Aoike (Research and Development for Next-Generation Systems Office, National Diet Library)
11:25-11:50 市民参加とAI―「みんなで翻刻」開発者の立場から Yuta Hashimoto (National Museum of Japanese History)
11:50-13:10 Break / Demonstration by speakers
13:10-13:15 基調講演紹介 Asanobu Kitamoto (ROIS-DS Center for Open Data in the Humanities / National Institute of Informatics)
Keynote talks - Prospect of kuzushiji recognition by AI
13:15-13:50 文字認識研究の過去・現在・未来 Seiichi Uchida (Graduate School and Faculty of Information Science and Electrical Engineering, Kyushu University)
13:50-14:25 過去からの挑戦状-くずし字認識の未来- Takahiro Sasaki (Keio Institute of Oriental Classics)
Session 2 - Release of kuzushiji dataset with one million characters and digital humanities research
14:25-14:40 NIJL-NWプロジェクトとくずし字データセット Kazuaki Yamamoto (National Institute of Japanese Literature)
14:40-14:55 デジタル人文学研究とAIくずし字認識 Asanobu Kitamoto (ROIS-DS Center for Open Data in the Humanities / National Institute of Informatics)
14:55-15:25 Break / Demonstration by speakers
Session 3 - Kaggle competition
15:25-15:40 Kuzushiji competition for gathering ideas worldwide Tarin Clanuwat (ROIS-DS Center for Open Data in the Humanities / National Institute of Informatics)
15:40-15:55 What we learned from AI researchers worldwide tackling kuzushiji Alex Lamb (MILA, University of Montreal)
Session 4 - Kaggle award ceremony, winner talks, and discussion
- Award ceremony Robert Campbell (National Institute of Japanese Literature)
Masaru Kitsuregawa (National Institute of Informatics)
- Winner talks 2nd Konstantin Lopukhin (Online)
3rd Kenji Doi (Venue)
5th See-- (Online)
7th K_mat (Venue)
8th Toshinori Hanya (Venue)
9th Marcos (Venue)
- Discussion Robert Campbell (National Institute of Japanese Literature) and others
17:40-17:45 Closing remarks Kazuaki Yamamoto (National Institute of Japanese Literature)

Posters

  1. Tarin Clanuwat, Alex Lamb, くずし字データセット・ KuroNetデモ
  2. Chikahiko Suzuki, Yingtao Tian, 顔貌コレクション(顔コレ)と顔コレデータセット
  3. Le Duc Anh, Deep Learning for Recognition of Modern Japanese Magazines
  4. Mika Ichino, 過去の記録を統合解析し、超学際的な課題に挑む歴史ビッグデータ

Releases

  1. 「日本文化とAIシンポジウム2019~AIがくずし字を読む時代がやってきた~」11月11日に開催、参加申込スタート
  2. 「日本文化とAIシンポジウム2019 ~AIがくずし字を読む時代がやってきた~」  11月11日に開催、参加申込スタート

Symposium Flyer

Plaese feel free to download and use the symposium flyer.