Kawahara Lab at Department of Communications and Computer Engineering, Waseda University

Kawahara Lab at Department of Communications and Computer Engineering,
Waseda University

About

We study and develop artificial intelligence systems that can understand language like humans.
In particular, by combining findings and technologies of computer science and linguistics, we work on world knowledge acquisition/utilization and text analysis/understanding.
Furthermore, we aim at elucidating the mechanisms by which humans understand language and developing systems which incorporate such mechanisms.

News

March 4, 2024:Open Lab 2024: We will have an open lab on March 18-21. If you are interested in being assigned to Kawahara Lab, please join us by making an appointment from here.

October 23, 2023:A paper at IJCNLP-AACL 2023 Student Research Workshop (SRW)

We will present the following paper at IJCNLP-AACL 2023 Student Research Workshop (November 1).

Tomohito Kasahara and Daisuke Kawahara. Exploring Automatic Evaluation Methods based on a Decoder-based LLM for Text Generation.

July 23, 2023: A paper at ACL 2023 Student Research Workshop (SRW)

We will present the following paper at ACL2023 Student Research Workshop (July 10-12).

Sakura Imai, Daisuke Kawahara, Naho Orita and Hiromune Oda. Theoretical Linguistics Rivals Embeddings in Language Clustering for Multilingual Named Entity Recognition.

June 26, 2023: Released BigBird (base) pre-trained model for Japanese

BigBird (base) model pre-trained on a dump of tokenized (pre-processed) Japanese is released via the Huggingface Models hub.

nlp-waseda/bigbird-base-japanese

May 2, 2023: Acceptance to ACL 2023

Our paper was accepted to Findings of ACL2023 (July 9-14)

Hao Wang, Hirofumi Shimizu, Daisuke Kawahara. Kanbun-LM: Reading and Translating Classical Chinese in Japanese Methods by Language Models.

April 1, 2023: New members: 9 students (7 bachelors, 1 master, 1 doctor) joined our lab.

March 7, 2023: Japanese ATOMIC and COMET released

The Japanese versions of ATOMIC, a commonsense knowledge graph about events, and COMET, a commonsense generation model, constructed in a joint research project with LINE Corporation, are now available.

GitHub Repo

March 4, 2023: Ski/Snowboard Camp: We held a ski/snowboard camp at Inawashiro Ski Resort from March 2 to 4.

February 28, 2023: Open Lab 2023: We will have an open lab on March 20-23. If you are interested in being assigned to Kawahara Lab, please join us by making an appointment from here.

September 24, 2022: Summer Seminar Camp: A summer seminar camp was held at the Karuizawa Seminar House from 22-24 September.

June 23, 2022: 3 papers at NAACL 2022 Student Research Workshop (SRW)

We will present the following papers at NAACL 2022 SRW (July 10-15).

Ryoma Sakaeda and Daisuke Kawahara. Generate, Evaluate, and Select: A Dialogue System with a Response Evaluator for Diversity-Aware Response Generation.
Ritvik Choudhary and Daisuke Kawahara. Grounding in Social Media: An Approach to Building a Chit-chat Dialogue Model.
Tomohito Kasahara, Daisuke Kawahara, Nguyen Tung, Shengzhe Li, Kenta Shinzato, and Toshinori Sato. Building a Personalized Dialogue System with Prompt-Tuning.

April 24, 2022: A paper at ACL 2022 Student Research Workshop (SRW)

We will present the following paper at ACL 2022 SRW (May 22-27).

Tatsuya Ide and Daisuke Kawahara. Building a Dialogue Corpus Annotated with Expressed and Experienced Emotions.

April 11, 2022: A paper at LREC 2022

We will present the following paper at LREC 2022 (June 21-25).

Kentaro Kurihara, Daisuke Kawahara, and Tomohide Shibata. JGLUE: Japanese General Language Understanding Evaluation.

April 1, 2022: New members.: 6 students joined our lab.

February 14, 2022: Open Lab 2022: We will have an open lab on March 22-24 in a hybrid manner. If you are interested in being assigned to Kawahara Lab, please join us by making an appointment from here.

December 29, 2021: Released GPT-2 (small) pre-trained model for Japanese

GPT-2 (small) model pre-trained solely on a dump of tokenized (pre-processed) Japanese Wikipedia is released via the Huggingface Models hub.

nlp-waseda/gpt2-small-japanese-wikipedia

December 22, 2021: Released RoBERTa (base) pre-trained model for Japanese

RoBERTa (base) model pre-trained on tokenized (pre-processed) Japanese text is released via the Huggingface Models hub.

nlp-waseda/roberta-base-japanese

August 3, 2021: Joint Research Project with LINE Corporation: We have started a joint research project with LINE Corporation, "Research on dialogue systems using ultra-large-scale Japanese language models". We will conduct research on how to effectively use and evaluate ultra-large-scale Japanese language models.

July 15, 2021: Joint Research Project with Yahoo Japan Corporation: We have started a joint research project with Yahoo Japan Corporation, "Construction of a Japanese evaluation set in natural language processing". We plan to build and publish a Japanese language understanding evaluation benchmark.

April 1, 2021: New members: 7 students joined our lab.

March 27, 2021: A paper at NAACL-HLT 2021 Student Research Workshop (SRW)

We will present the following paper at NAACL-HLT 2021 SRW (June 6-11).

Tatsuya Ide and Daisuke Kawahara. Multi-Task Learning of Generation and Classification for Emotion-Aware Dialogue Response Generation.

January 15, 2021: Appointment for Lab Assignment 2021: If you want to have a meeting to discuss the assignment to Kawahara Lab on March 22 or 23, 2021, please make an appointment at this form. The location is online. If you have any questions, please send an email to the mail address below.

July 15, 2020: Worldwide COVID-19 Information Aggregation Site: In collaboration with computer science laboratories of universities and research institutes, we have developed a site that aggregates worldwide information on COVID-19.

April 1, 2020: Beginning of Research Activity: We will have the first meeting with the assigned students on April 1 online and offline.

March 14, 2020: Appointment for Lab Assignment: If you want to have a meeting to discuss the assignment to Kawahara Lab on March 20 or 23, 2020, please make an appointment at this form. The location is the Multimedia Room at the 9th floor, Bldg. 55 N. If you have any questions, please send an email to the mail address below.