Prof. Shirakawa has been selected as an Outstanding Reviewer (Top 10%) at ICML 2022. [Link]

Our paper has been accepted to the 31st International Conference on Artificial Neural Networks (ICANN 2022). This paper proposes an efficient search method for multiple neural architectures with different complexities by using importance sampling.

• Yuhei Noda, Shota Saito, and Shinichi Shirakawa: Efficient Search of Multiple Neural Architectures with Different Complexities via Importance Sampling, 31st International Conference on Artificial Neural Networks (ICANN 2022), September 6-9 2022 (Accepted). [arXiv]

Our papers (two full papers, one poster paper, and one workshop paper) have been accepted to the Genetic and Evolutionary Computation Conference (GECCO) 2022.

The members’ page has been updated. Now, our laboratory has 5 doctoral course students, 14 master’s course students, and 6 undergraduate students for graduation research.

[Members page]

Our paper regarding text-to-gesture generation based on deep learning has been accepted to Neural Networks. This work is based on the master’s course research by Asakawa and is collaborative research with Prof. Hasegawa at Hokkai Gakuen University and Prof. Kaneko at Aoyama Gakuin University.

Eiichi Asakawa, Naoshi Kaneko, Dai Hasegawa, and Shinichi Shirakawa: Evaluation of text-to-gesture generation model using convolutional neural network, Neural Networks, Elsevier [Link]

The following paper has been published in Scientific Reports. This work is the collaborative research with Prof. Fukuda and Prof. Ohmori at Yokohama National University.

Minami Masumoto, Ittetsu Fukuda, Suguru Furihata, Takahiro Arai, Tatsuto Kageyama, Kiyomi Ohmori, Shinichi Shirakawa, and Junji Fukuda: Deep neural network for the determination of transformed foci in Bhas 42 cell transformation assay, Scientific Reports, volume 11, Article number: 23344, Dec. 2021. [Link]

The following paper has been accepted to the 13th Asian Conference on Machine Learning (ACML 2021). This paper provides a benchmark dataset for joint optimization of architecture and training hyperparameters. The benchmark dataset API is available from the GitHub repository.

Yoichi Hirose, Nozomu Yoshinari, and Shinichi Shirakawa, “NAS-HPO-Bench-II: A Benchmark Dataset on Joint Optimization of Convolutional Neural Network Architecture and Training Hyperparameters,” The 13th Asian Conference on Machine Learning (ACML 2021) (Accepted) [arXiv]

The members’ page has been updated. Now, our laboratory has 4 doctoral course students, 14 master’s course students, and 5 undergraduate students.

[Members page]

The following paper has been accepted to the Computer Graphics International (CGI) 2020 as a proceedings paper. This paper proposes an improved redirected walking technique using reinforcement learning, where the redirected walking enables to experience virtual reality in limited physical spaces.

Wataru Shibayama and Shinichi Shirakawa: Reinforcement Learning-Based Redirection Controller for Efficient Redirected Walking in Virtual Maze Environment, Computer Graphics International (CGI) 2020 (Accepted)

The following paper has been accepted to the The Sixteenth International Conference on Parallel Problem Solving from Nature (PPSN 2020). This paper improves the adaptive stochastic natural gradient method (ASNG) to work it well on the objective functions with low effective dimensionality.

Teppei Yamaguchi, Kento Uchida, and Shinichi Shirakawa: Adaptive Stochastic Natural Gradient Method for Optimizing Functions with Low Effective Dimensionality, The Sixteenth International Conference on Parallel Problem Solving from Nature (PPSN 2020) (Accepted)