I'm Tatsuya Konishi (Japanese). I hold a Master's degree from the Graduate School of Information Science and Technology, The University of Tokyo. I'm working at KDDI Research, Inc., a research institute of KDDI Corporation, one of Japan's major telecommunications companies. I have also joined C3F in which developers are contributing to the Internet Computer ecosystem.

My tech-interest includes:

• Research
  • Continual learning
  • Graph neural networks
  • Behavior change
• Engineering
  • Android apps
  • Crypto currencies
    • Internet Computer ($ICP)

• Website (Japanese)
• Twitter
• Facebook
• LinkedIn
• Google Scholar

• SPG (accepted in ICML2023)
  • Interview article on Nikkei Robotics

• TBA

• dependroid (>10k installs)
• Oh my shortcut (>500 installs)
• TBA

• HR Technology Awards 2020
  • Award for Excellence in Human Resources Management.
  • Official page
  • Company's news release
• IoT Innovation Challenge 2018
  • Finalist (4 out of 19 teams)
  • Official reports
  • Presentation
• Irregular Congestion Prediction (done in research internship at Yahoo! JAPAN R&D during 2016-2017)
  • Research results are now available in the app.
  • Official page
  • Company's news release

As a first author:

• Tatsuya Konishi, Mori Kurokawa, Chihiro Ono, Zixuan Ke, Gyuhak Kim, and Bing Liu. Parameter-Level Soft-Masking for Continual Learning. In Proc. of ICML, 2023.
• Tatsuya Konishi*, Shuichiro Haruta*, Mori Kurokawa, Kenta Tsukatsune, Yuto Mizutani, Tomoaki Saito, Hideki Asho, and Chihiro Ono. CG-GNN: A Novel Compiled Graphs-based Feature Extraction Method for Enterprise Social Networks. In Proc. of ICDAR, 2023.
• Tatsuya Konishi, Mori Kurokawa, Chihiro Ono, Zixuan Ke, Gyuhak Kim, and Bing Liu. Partially Relaxed Masks for Knowledge Transfer Without Forgetting in Continual Learning. In Proc. of PAKDD, 2022.
• Tatsuya Konishi, Masatoshi Nagata, Masaru Honjo, Akio Yoneyama, Masayuki Kurokawa, and Koji Mishima. Social Behaviors: A Social Topology and Interaction Pattern Affect the Properties of a Changed Behavior. In Proc. of Persuasive Technology, 2019.
• Tatsuya Konishi, Mikiya Maruyama, Kota Tsubouchi, and Masamichi Shimosaka. Cityprophet: City-scale irregularity prediction using transit app logs. In Proc. of UbiComp, 2016.

As a co-author:

• Gyuhak Kim, Changnan Xiao, Tatsuya Konishi, and Bing Liu. Learnability and Algorithm for Continual Learning. In Proc. of ICML, 2023.
• Zixuan Ke, Yijia Shao, Haowei Lin, Tatsuya Konishi, Gyuhak Kim, and Bing Liu. Continual Pre-training of Language Models. In Proc. of ICLR, 2023.
• Shuichiro Haruta, Tatsuya Konishi, and Mori Kurokawa. A Novel Graph Aggregation Method Based on Feature Distribution Around Each Ego-node for Heterophily. In Proc. of ACML, 2023.
• Gyuhak Kim, Changnan Xiao, Tatsuya Konishi, Zixuan Ke, and Bing Liu. A Theoretical Study on Solving Continual Learning. In Proc. of NeurIPS, 2022.
• Roberto Legaspi, Wenzhen Xu, Tatsuya Konishi, Shinya Wada, and Yuichi Ishikawa. Multidimensional Analysis of Sense of Agency During Goal Pursuit. In Proc. of UMAP, 2022.
• Roberto Legaspi, Wenzhen Xu, Tatsuya Konishi, and Shinya Wada. Positing a Sense of Agency-Aware Persuasive AI: Its Theoretical and Computational Frameworks. In Proc. of Persuasive Technology, 2021.
• Kenta Tsukatsune, Tatsuya Konishi, Yuto Mizutani, Mori Kurokawa, Shuichiro Haruta, and Tomoaki Saito. Does Activeness Originate in Individuals or Groups? Analysis of the Interrelationships between Network Indices and Posts Using Multilevel Cross-Lagged Model. In Proc. of WI-IAT, 2021.
• Ryo Imai, Kota Tsubouchi, Tatsuya Konishi, and Masamichi Shimosaka. Early Destination Prediction with Spatio-temporal User Behavior Patterns. In Proc. of IMWUT, 2018.

• JP.2023064200.A
• JP.7279225.B2, US.2023086727.A1
• JP.2023030306.A
• JP.2022154041.A
• JP.7237895.B2
• JP.2022016277.A
• JP.7282052.B2
• JP.7296905.B2
• JP.7328162.B2