Research

Journal Papers

  1. Muneki Yasuda and Kaiji Sekimoto: Gaussian-Discrete restricted Boltzmann machine with Sparse-Regularized Hidden Layer, Behaviormetrika, 2024.
  2. Kaiji Sekimoto, Chako Takahashi, and Muneki Yasuda: Quasi-free Energy Evaluation of Gaussian-Bernoulli Restricted Boltzmann Machine for Anomaly Detection, Nonlinear Theory and its Applications (NOLTA), IEICE, Vol. 15, No. 2, pp. 273-283, 2024.
  3. Kaiji Sekimoto and Muneki Yasuda: Effective learning algorithm for restricted Boltzmann machines via spatial Monte Carlo integration, Nonlinear Theory and its Applications (NOLTA), IEICE, Vol. 14, No. 2, pp. 228-241, 2023.
  4. Kaiji Sekimoto and Muneki Yasuda: Composite Spatial Monte Carlo Integration Based on Generalized Least Squares, Journal of the Physical Society of Japan, Vol. 91, No. 11, Article ID. 114003, 2022.
  5. Muneki Yasuda and Kaiji Sekimoto: Spatial Monte Carlo integration with annealed importance sampling, Physical Review E, Vol. 103, No. 5, Article No. 052118, 2021.

International Conferences

  1. Chako Takahashi, Kaiji Sekimoto, and Muneki Yasuda: Free energy evaluation based on Bethe free energy on slim deep Boltzmann machines, Proceedings of the 2024 International Symposium on Nonlinear Theory and its Applications (NOLTA2024), accepted.
  2. Kaiji Sekimoto, Chako Takahashi, and Muneki Yasuda: Quasi-Free Energy Evaluation of Restricted Boltzmann Machine for Anomaly Detection, Proceedings of the 2023 International Symposium on Nonlinear Theory and its Applications (NOLTA2023), pp. 142-145, 2023.
  3. Kaiji Sekimoto and Muneki Yasuda: Spatial Monte Carlo Integration for Learning Restricted Boltzmann Machine, Proceedings of the 2022 International Symposium on Nonlinear Theory and its Applications (NOLTA2022), pp. 9-12, 2022. [Student Paper Award]

Domestic Conferences

  1. Kaiji Sekimoto and Muneki Yasuda: Efficient Maximum Marginal Probability Estimation on Restricted Boltzmann Machines, The 27th Information-Based Induction Sciences Workshop (IBIS2024), 2-R-034, Sonic City (Saitama), November 6, 2024.
  2. Kaiji Sekimoto and Muneki Yasuda: Improvement of Incomplete Data Learning in Gaussian--Bernoulli Restricted Boltzmann Machine, The 23th Forum on Information Technology (FIT2024), CF-007, Hiroshima Institute of Technology, September 4, 2024.
  3. Ryusuke Ishioka, Kaiji Sekimoto, and Muneki Yasuda: Statistical analysis of combinatorial optimization problems based on hierarchical Bayesian learning, The 86th National Convention of IPSJ, 4L-06, Kanagawa University (Hybrid Meeting), March 16, 2024.
  4. Tomoki Haga, Kaiji Sekimoto, and Muneki Yasuda: Markov random field with TV-regularization-type interactions, The 86th National Convention of IPSJ, 2L-09, Kanagawa University (Hybrid Meeting), March 15, 2024.
  5. Kaiji Sekimoto and Muneki Yasuda: Restricted-Boltzmann-machine learning using incomplete dataset, The 85th National Convention of IPSJ, 4M-01, The University of Electro-Communications (Hybrid Meeting), March 3, 2023. [Student Encouragement Award]
  6. Kaiji Sekimoto and Muneki Yasuda: Composite Spatial Monte Carlo Integration Based on Generalized Least Squares, The Physical Society of Japan 2022 Annual (77th) Meeting, 15aB14-12, Online Meeting, March 5, 2022.
  7. Kaiji Sekimoto and Muneki Yasuda: Composite Spatial Monte Carlo Integration Based on Generalized Least Squares, The 84th National Convention of IPSJ, 7M-04, Ehime University (Hybrid Meeting), March 5, 2022. [Student Encouragement Award]
  8. Kaiji Sekimoto and Muneki Yasuda: Effective sampling approximation based on spatial Monte Carlo integration and Annealed importance sampling, The 83th National Convention of IPSJ, 1J-01, Online Meeting, March 18, 2021. [Student Encouragement Award]