Publications

International Conferences

1. Flowchart-Based Decision Making with Large Language Models,
   Yuuki Yamanaka, Hiroshi Takahashi, Tomoya Yamashita,
   Findings of ACL, 2025.
   [paper]
2. Positive-unlabeled AUC Maximization under Covariate Shift,
   Atsutoshi Kumagai, Tomoharu Iwata, Hiroshi Takahashi, Taishi Nishiyama, Kazuki Adachi, Yasuhiro Fujiwara,
   ICML, 2025.
   [paper]
3. Positive-Unlabeled Diffusion Models for Preventing Sensitive Data Generation,
   Hiroshi Takahashi, Tomoharu Iwata, Atsutoshi Kumagai, Yuuki Yamanaka, Tomoya Yamashita,
   ICLR, 2025.
   [paper] [arXiv] [code] [slides] [poster]
4. Importance-weighted Positive-unlabeled Learning for Distribution Shift Adaptation,
   Atsutoshi Kumagai, Tomoharu Iwata, Hiroshi Takahashi, Taishi Nishiyama, Yasuhiro Fujiwara,
   AISTATS, 2025.
   [paper]
5. AUC Maximization under Positive Distribution Shift,
   Atsutoshi Kumagai, Tomoharu Iwata, Hiroshi Takahashi, Taishi Nishiyama, Yasuhiro Fujiwara,
   NeurIPS, 2024.
   [paper]
6. One-vs-the-Rest Loss to Focus on Important Samples in Adversarial Training,
   Sekitoshi Kanai, Shin’ya Yamaguchi, Masanori Yamada, Hiroshi Takahashi, Yasutoshi Ida,
   ICML, 2023.
   [paper] [arXiv]
7. Meta-learning for Robust Anomaly Detection,
   Atsutoshi Kumagai, Tomoharu Iwata, Hiroshi Takahashi, Yasuhiro Fujiwara,
   AISTATS, 2023.
   [paper]
8. Learning Optimal Priors for Task-Invariant Representations in Variational Autoencoders,
   Hiroshi Takahashi, Tomoharu Iwata, Atsutoshi Kumagai, Sekitoshi Kanai, Masanori Yamada, Yuuki Yamanaka, Hisashi Kashima,
   KDD, 2022.
   [paper] [slides] [slides (Japanese)] [poster] [poster (Japanese)]
9. Constraining Logits by Bounded Function for Adversarial Robustness,
   Sekitoshi Kanai, Masanori Yamada, Shin’ya Yamaguchi, Hiroshi Takahashi, Yasutoshi Ida,
   IJCNN, 2021.
   [paper] [arXiv]
10. Autoencoding Binary Classifiers for Supervised Anomaly Detection,
    Yuki Yamanaka, Tomoharu Iwata, Hiroshi Takahashi, Masanori Yamada, Sekitoshi Kanai,
    PRICAI, 2019.
    [paper] [arXiv]
11. Variational Autoencoder with Implicit Optimal Priors,
    Hiroshi Takahashi, Tomoharu Iwata, Yuki Yamanaka, Masanori Yamada, Satoshi Yagi,
    AAAI, 2019.
    [paper] [arXiv] [code] [slides] [poster]
12. Student-t Variational Autoencoder for Robust Density Estimation,
    Hiroshi Takahashi, Tomoharu Iwata, Yuki Yamanaka, Masanori Yamada, Satoshi Yagi,
    IJCAI, 2018.
    [paper] [code] [slides]

Journal Articles

1. Deep Positive-Unlabeled Anomaly Detection for Contaminated Unlabeled Data,
   Hiroshi Takahashi, Tomoharu Iwata, Atsutoshi Kumagai, Yuuki Yamanaka,
   Neurocomputing, 2026, 135134.
   [paper] [arXiv] [code] [poster (Japanese)]
2. Relationship Between Nonsmoothness in Adversarial Training, Constraints of Attacks, and Flatness in the Input Space,
   Sekitoshi Kanai, Masanori Yamada, Hiroshi Takahashi, Yuki Yamanaka, Yasutoshi Ida,
   IEEE Transactions on Neural Networks and Learning Systems, 2024, 35(8), 10817–10831.
   [paper]
3. Student-t Variational Autoencoder for Robust Multivariate Density Estimation,
   Hiroshi Takahashi, Tomoharu Iwata, Yuki Yamanaka, Masanori Yamada, Satoshi Yagi, Hisashi Kashima,
   Transactions of the Japanese Society for Artificial Intelligence, 2021, 36(3), A-KA4_1–9. (in Japanese)
   [paper]

Preprints

1. Relative Density Ratio Optimization for Stable and Statistically Consistent Model Alignment,
   Hiroshi Takahashi, Tomoharu Iwata, Atsutoshi Kumagai, Sekitoshi Kanai, Masanori Yamada, Kosuke Nishida, Kazutoshi Shinoda,
   arXiv:2604.04410, 2026.
   [arXiv] [code]
2. Test-Time Alignment of LLMs via Sampling-Based Optimal Control in pre-logit space,
   Sekitoshi Kanai, Tsukasa Yoshida, Hiroshi Takahashi, Haru Kuroki, Kazumune Hashimoto,
   arXiv:2510.26219, 2025.
   [arXiv]
3. Smoothness Analysis of Loss Functions of Adversarial Training,
   Sekitoshi Kanai, Masanori Yamada, Hiroshi Takahashi, Yuki Yamanaka, Yasutoshi Ida,
   arXiv:2103.01400, 2021.
   [arXiv]
4. Adversarial Training Makes Weight Loss Landscape Sharper in Logistic Regression,
   Masanori Yamada, Sekitoshi Kanai, Tomoharu Iwata, Tomokatsu Takahashi, Yuki Yamanaka, Hiroshi Takahashi, Atsutoshi Kumagai,
   arXiv:2102.02950, 2021.
   [arXiv]

Thesis

• Improving Variational Autoencoders on Robustness, Regularization, and Task-Invariance,
  Hiroshi Takahashi,
  Doctoral Thesis, Graduate School of Informatics, Kyoto University, 2023.
  [paper]

Lectures

1. Generative Models: Foundations and Applications (2025)
   Hiroshi Takahashi,
   Graduate School of Information Science and Technology, Osaka University, June 5, 2025.
   [slides (Japanese)]
2. Generative Models: Foundations and Applications (2024)
   Hiroshi Takahashi,
   Graduate School of Information Science and Technology, Osaka University, June 13, 2024.
   [slides (Japanese)]