Privacy-Preserving Data Analysis and Secure Data Infrastructure for Real Applications

Japanese / English

Objectives

Our goal is to advance research in system software (data infrastructure, OS), data analysis (differential privacy, federated learning), and real-world applications (medical data, trajectory data) to alleviate concerns such as data leakage due to unauthorized accesses, innocent but leaky data extraction by legitimate users, and the identification of individuals (intentional or unintentional) from analysis results. We aim to contribute to the realization of a Society 5.0 that allows safe and proactive data utilization.

This project has been adopted in the JST CREST field “Creation of System Software for Society 5.0 by Integrating Fundamental Theories and System Platform Technologies”.

Team

• Principal Investigator
  • Kenjiro Taura - The University of Tokyo Graduate School of Information Science and Technology, Department of Electronic Information
• Principal Research Collaborators
  • Masatoshi Yoshikawa - Kyoto University Graduate School of Informatics, Department of Social Informatics
  • Shouhei Hanaoka - The University of Tokyo Hospital, Department of Radiology, Image Information Processing and Analysis Laboratory
  • Toshihiro Hanawa - The University of Tokyo Information Technology Center
  • Yang Cao - Hokkaido University Institute of Information Science, Division of Information and Computer Science
• Collaborators
  • Kenjiro Taura Group
    • Osamu Tatebe
    • JIANG Renhe
    • Kento Aida
  • Shouhei Hanaoka Group
    • Hisaichi Shibata - DP-Glow
    • Tomomi Takenaga - Local differential privacy protection for multi-dimensional medical checkup data: real-world validation (under review)
    • Alam MD Ashraful - Estimation of future occurrence of HbA1c elevation with and without Differential Privacy (under review)
    • Tomohiro Kikuchi @ Jichi Medical University Radiology Department - Synthetic data generation method for hybrid image-tabular data using two generative adversarial networks.
    • Yukihiro Nomura @ Chiba University Nomura Laboratory - DPSGD
    • Aiki Yamada @ Chiba University Nomura Laboratory - Investigation of federated learning for automated cerebral aneurysm detection in head MR angiography images (presented in CARS 2023)

Research Topics

• Many valuable data sources, such as medical data (diagnostic images, electronic health records, prescription data) and location data of individuals and objects, are based on personal information.
• Typically, such data is provided after removing personally identifiable information (pseudonymization, hashing) and used for purposes such as building machine learning models (e.g., constructing diagnostic models using a large number of normal and pathological images) and performing statistical calculations (e.g., calculating the number of users per period or per facility) without an intent to expose personal information.
• However, it is well-known that even in such cases, it is still possible to identify individual information by combining the analysis results (machine learning models or statistical data) with external knowledge about individuals, which poses a barrier to providing and using personal information for legitimate purposes.
• Additionally, current file systems and cloud storage solutions equate providing data and granting read access to data, making it difficult for systems to control the copying and leakage of read data. This also acts as a barrier to data provision.
• To address these issues, new system technologies are required, such as systems that limit the extent to which read data can leak and/or systems that prohibit leaking results that can identify individuals.
• In addition to these approaches, there are methods to anonymize data itself more thoroughly in such a way that it is impossible to recover personal information (e.g., K-anonymization and local differential privacy), but these methods come with the problem of significantly reducing the utility of the data.
• In this research, we address these issues at various levels (foundation software, algorithmic frameworks, practical applications) and collaborate on the following modules (all ongoing research with the possibility of changes in goals and approaches):

1. Secure File System that Do Not Rely on Trustworthy Administrators
2. System Mechanisms for Enforcing and Traceable Privacy Protection
3. Flexible Privacy-Preserving Data Analysis and Machine Learning
4. Demonstrations in Medical and Trajectory Data Applications

Secure File System that Do Not Rely on Trustworthy Administrators

System Mechanisms for Enforcing and Traceable Privacy Protection

• In this sub-theme, we research systems, specifically programming language processing systems, to prevent the output of analysis results that can identify personal information. We utilize the framework of (global) differential privacy to ensure that personal information cannot be identified from the analysis results.
• In the framework of global differential privacy, stochastic noise is added to the analysis results (e.g., averages or sums of a large number of individual data) to make the contribution of specific individuals to the results ambiguous, preventing the inference of individual data from the results.
• In the target language processing system:
  • Data based on personal information (inputs from files or databases) and data derived from calculations based on it are prohibited from being output.
  • Only data processed through appropriate global differential privacy mechanisms (with added stochastic noise) are allowed to be output, ensuring privacy protection.

Flexible Privacy-Preserving Data Analysis and Machine Learning

In this sub-theme, we research the following topics:

• Privacy-Preserving Federated Learning Using Trusted Execution Environments (TEEs)

Combining federated learning (FL) with trusted execution environments (TEEs) is a promising approach to achieve privacy-preserving FL. Server-side TEE implementation enables FL to progress without exposing gradient information to the server during each round, particularly benefiting utility in FL using local differential privacy. However, considering the vulnerability of server-side TEE in the context of FL has not been sufficiently researched. In this sub-topic, we design systems and algorithms to analyze the vulnerability of TEE in FL, propose rigorous defense methods to enhance TEE, and develop systems that can utilize TEE for private and high-utility federated learning.

• Privacy-Preserving Contribution Computation in Federated Learning

Evaluating the contribution of each client’s data is crucial in federated learning for applications such as data markets, explainable AI, or malicious client detection. Shapley value (SV) is a commonly used metric for contribution evaluation. However, existing methods for calculating SV in FL do not consider privacy. They assume that the server can access both unencrypted FL models and client data without privacy protection. Therefore, in this topic, we research privacy-preserving SV computation. We propose a protocol for efficient and private SV calculation, SecSV, for cross-silo FL. SecSV uses a hybrid privacy protection scheme to avoid ciphertext-ciphertext multiplication between test data and models. Experimental results show that SecSV is 5.9 to 18.0 times faster than the Baseline using homomorphic encryption.

Demonstrations in Medical and Trajectory Data Applications

1. Medical Sub-Theme
   • In the medical sub-theme, we primarily conduct three demonstrations:
   • Development, deployment, and practical use of privacy protection systems using differential privacy in medical settings.
   • Collaboration with Kenjiro Taura
   • Collaboration with Toshihiro Hanawa
   • Collaboration with Yang Cao

Here, the main goal is to establish AI learning methods that protect patient information used in AI learning while using the privacy enforcement technology developed by Kenjiro Taura. AI learning is carried out using the framework of differentially private stochastic gradient descent (DPSGD) to minimize the risk of patient information leakage.

• Collaboration with Fumiharu Kato

In this collaboration, the focus is on the deployment side, specifically how to safely and scalably execute pre-trained AI and deliver the results to clinical practitioners and patients using secure computation and supercomputers.

Publications

1. Shun Takagi, Fumiharu Kato, Yang Cao and Masatoshi Yoshikawa. “Asymmetric Differential Privacy.” 2022 IEEE International Conference on Big Data (Big Data). 2022.
2. Ryota Hiraishi, Masatoshi Yoshikawa, Shun Takagi, Yang Cao, Sumio Fujita and Hidehito Gomi. “Mitigating Privacy Vulnerability Caused by Map Asymmetry.” Lecture notes in computer science. 2022.
3. Shuyuan Zheng, Yang Cao, Masatoshi Yoshikawa, Huizhong Li and Yan Qiang. “FL-Market: Trading Private Models in Federated Learning.” 2022 IEEE International Conference on Big Data (Big Data). 2022.
4. Shang Liu, Yang Cao, Takao Murakami and Masatoshi Yoshikawa. “A Crypto-Assisted Approach for Publishing Graph Statistics with Node Local Differential Privacy.” 2022 IEEE International Conference on Big Data (Big Data). 2022.
5. 空閑 洋平 and 中村 遼. “遠隔会議システムの計測データを用いた広域ネットワーク品質計測.” インターネットと運用技術シンポジウム論文集 . 2022.
6. Ruixuan Cao, Fumiharu Kato, Yang Cao and Masatoshi Yoshikawa. “An Accurate, Flexible and Private Trajectory-Based Contact Tracing System on Untrusted Servers.” Lecture notes in computer science. 2022.
7. Shun Takagi, Fumiharu Kato, Yang Cao and Masatoshi Yoshikawa. “From Bounded to Unbounded: Privacy Amplification via Shuffling with Dummies.” 2023 IEEE 36th Computer Security Foundations Symposium (CSF). 2023.
8. Shuyuan Zheng, Yang Cao and Masatoshi Yoshikawa. “Secure Shapley Value for Cross-Silo Federated Learning.” Proceedings of the VLDB Endowment. 2023.
9. Hisaichi Shibata, Shouhei Hanaoka, Yang Cao, Masatoshi Yoshikawa, Tomomi Takenaga, Yukihiro Nomura, Naoto Hayashi and Osamu Abe. “Local Differential Privacy Image Generation Using Flow-Based Deep Generative Models.” Applied sciences. 2023.
10. Shumpei Shiina and Kenjiro Taura. “Itoyori: Reconciling Global Address Space and Global Fork-Join Task Parallelism.” In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis (SC ‘23). 2023.
11. Chao Tan, Yang Cao, Sheng Li and Masatoshi Yoshikawa. “General or Specific? Investigating Effective Privacy Protection in Federated Learning for Speech Emotion Recognition.” ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). 2023.
12. Shun Takagi, Yang Cao, Yasuhito Asano and Masatoshi Yoshikawa. “Geo-Graph-Indistinguishability: Location Privacy on Road Networks with Differential Privacy.” IEICE transactions on information and systems. 2023.
13. Ruixuan Liu, Yang Cao, Yanlin Wang, Lingjuan Lyu, Yun Chen and Hong Chen. “PrivateRec: Differentially Private Model Training and Online Serving for Federated News Recommendation.” KDD ‘23: Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. 2023.
14. Xuyang Cao, Yang Cao, Primal Pappachan, Atsuyoshi Nakamura and Masatoshi Yoshikawa. “Differentially Private Streaming Data Release Under Temporal Correlations via Post-processing.” Lecture notes in computer science. 2023.
15. Xiaoyu Li, Yang Cao and Masatoshi Yoshikawa. “Locally Private Streaming Data Release with Shuffling and Subsampling.” 2023 IEEE 39th International Conference on Data Engineering Workshops (ICDEW). 2023.
16. Fumiharu Kato, Yang Cao and Masayuki Yoshikawa. “Olive: Oblivious Federated Learning on Trusted Execution Environment against the Risk of Sparsification.” Proceedings of the VLDB Endowment. 2023.
17. Ryota Hiraishi, Masatoshi Yoshikawa, Yang Cao, Sumio Fujita and Hidehito Gomi. “Mechanisms to Address Different Privacy Requirements for Users and Locations.” IEICE transactions on information and systems. 2023.
18. Haotian Gao, Renhe Jiang, Zheng Dong, Jinliang Deng and Xuan Song. “Spatio-Temporal-Decoupled Masked Pre-training: Benchmarked on Traffic Forecasting.” Proc. of the 33rd International Joint Conference on Artificial Intelligence (IJCAI). 2023.
19. Yamada A, Hanaoka S, Shibata H, Takenaga T, Yoshikawa T and Nomura Y. “Investigation of federated learning for automated cerebral aneurysm detection in head MR angiography images.” CARS 2023 the 37th International Congress and Exhibition of Computer Assisted Radiology and Surgery, Munich, Germany June 20-23, 2023 (Int J Comput Assist Radio Surg (proc. CARS2023) 18(suppl.1): S63-S64)​. 2023.
20. 山田藍樹, 花岡昇平, 竹永智美, 柴田寿一, 吉川健啓 and 野村行弘. “病変自動検出における差分プライバシーを適用した連合学習手法の検討.” 第1回JAMIT若手医用画像工学シンポジウム: SAMIT2023, 茨城, 2023.9.30. 2023.
21. Fumiharu Kato, Li Xiong, Shun Takagi, Yang Cao and Masatoshi Yoshikawa. “ULDP-FL: Federated Learning with Across Silo User-Level Differential Privacy.” International Conference on Very Large Data Bases (VLDB). 2024.
22. Shuzo Takagi, Li Xiong, Fumiharu Kato, Yang Cao and Masatoshi Yoshikawa. “HRNet: Differentially Private Hierarchical and Multi-Resolution Network for Human Mobility Data Synthesization.” International Conference on Very Large Data Bases (VLDB). 2024.
23. Eric Chen, Yang Cao and Yifei Ge. “A Generalized Shuffle Framework for Privacy Amplification: Strengthening Privacy Guarantees and Enhancing Utility.” Proceedings of the … AAAI Conference on Artificial Intelligence. 2024.
24. . “Noise-Aware Algorithm for Heterogeneous Differentially Private Federated Learning.” Proceedings of the 41st International Conference on Machine Learning. 2024.
25. Eric Chen, Yang Cao and Yifei Ge. “Renyi Differential Privacy in the Shuffle Model: Enhanced Amplification Bounds.” IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP). 2024.
26. Hisaichi Shibata, Shouhei Hanaoka, Takahiro Nakao, Tomohiro Kikuchi, Y. Nakamura, Yukihiro Nomura, T. Yoshikawa and Osamu Abe. “Practical Medical Image Generation with Provable Privacy Protection Based on Denoising Diffusion Probabilistic Models for High-Resolution Volumetric Images.” Applied sciences. 2024.
27. Tomohiro Kikuchi, Shouhei Hanaoka, Takahiro Nakao, Tomomi Takenaga, Yukihiro Nomura, Harushi Mori and T. Yoshikawa. “Synthesis of Hybrid Data Consisting of Chest Radiographs and Tabular Clinical Records Using Dual Generative Models for COVID-19 Positive Cases.” Journal of Imaging Informatics in Medicine. 2024.
28. Shouhei Hanaoka, Yukihiro Nomura, T. Yoshikawa, Takahiro Nakao, Tomomi Takenaga, Hirotaka Matsuzaki, Nobutake Yamamichi and Osamu Abe. “Detection of pulmonary nodules in chest radiographs: novel cost function for effective network training with purely synthesized datasets.” International journal of computer assisted radiology and surgery. 2024.
29. 平岡拓海, 花岡昇平 and 田浦健次朗. “Pythonエコシステムに広範囲に適用可能な差分プライバシーのためのライブラリの実装と応用.” 電子情報通信学会 情報セキュリティ研究会. 2024.
30. Yamada A, Hanaoka S, Takenaga T, Yoshikawa T and Nomura Y. “Preliminary study of maximizing performance for each site in training of computer-aided diagnosis using federated learning.” CARS 2024 the 38th International Congress and Exhibition of Computer Assisted Radiology and Surgery, Barcelona, Spain, June 18–21, 2024 (Int J Comput Assist Radio Surg (proc. CARS2024) 19(suppl.1): S91-S92). 2024.
31. Alam MA, Hanaoka S, Nomura Y, Yoshikawa T and Abe O. “Local Differential Privacy for 3-D Images with Threshold Mechanism and Wavelet Transformation.” In: Proceedings of the 6th International Conference on Intelligent Medicine and Image Processing (IMIP 2024); 2024 Apr 24-26; Bali, Indonesia. 2024.
32. Lele Zheng, Yang Cao, Renhe Jiang, Kenjiro Taura, Yulong Shen, Sheng Li and Masatoshi Yoshikawa. “Enhancing Privacy of Spatiotemporal Federated Learning against Gradient Inversion Attacks.” The International Conference on Database Systems for Advanced Applications (DASFAA). 2024.
33. 山田 藍樹, 花岡 昇平, 竹永 智美, 吉川 健啓 and 野村 行弘. “パーソナライズド連合学習を用いたEOB-MR画像の肝結節性病変分類における施設毎の性能改善の検討.” 日本医用画像工学会大会. 2024.
34. Aiki Yamada, Shouhei Hanaoka, Tomomi Takenaga, Soichiro Miki, Takeharu Yoshikawa and Yukihiro Nomura. “Investigation of distributed learning for automated lesion detection in head MR images.” Radiological Physics and Technology. 2024.
35. Md Ashraful Alam, Shouhei Hanaoka, Yukihiro Nomura, Tomohiro Kikuchi, Takahiro Nakao, Tomomi Takenaga, Naoto Hayashi, T. Yoshikawa and Osamu Abe. “Improved identification of tumors in 18F-FDG-PET examination by normalizing the standard uptake in the liver based on blood test data.” International Journal of Computer Assisted Radiology and Surgery. 2024.
36. Hisaichi Shibata, Shouhei Hanaoka, Saori Koshino, Soichiro Miki, Yuki Sonoda and Osamu Abe. “Identity Diffuser: Preserving Abnormal Region of Interests while Diffusing Identity.” Applied Sciences. 2024.
37. Shouhei Hanaoka, Yukihiro Nomura, T. Yoshikawa, Takahiro Nakao, Tomomi Takenaga, Hirotaka Matsuzaki, Nobutake Yamamichi and Osamu Abe. “Detection of pulmonary nodules in chest radiographs: novel cost function for effective network training with purely synthesized datasets.” International Journal of Computer Assisted Radiology and Surgery. 2024.
38. Shouhei Hanaoka, Yukihiro Nomura, Naoto Hayashi, Issei Sato, Soichiro Miki, T. Yoshikawa, Hisaichi Shibata, Takahiro Nakao, Tomomi Takenaga, Hiroaki Koyama, Shinichi Cho, Noriko Kanemaru, Kotaro Fujimoto, Naoya Sakamoto, Tomoya Nishiyama, Hirotaka Matsuzaki, Nobutake Yamamichi and Osamu Abe. “Deep generative abnormal lesion emphasization validated by nine radiologists and 1000 chest X-rays with lung nodules.” PLoS ONE. 2024.
39. 山田藍樹, 花岡昇平, 竹永智美, 吉川健啓 and 野村行弘. “パーソナライズド連合学習を用いたEOB-MR 画像の肝結節性病変分類における施設毎の性能改善の検討.” 第43回日本医用画像工学会大会, OP6-5, 東京, 2024.8.5-7. 2024.
40. Aiki Yamada, Shouhei Hanaoka, Tomomi Takenaga, Soichiro Miki, T. Yoshikawa and Yukihiro Nomura. “Investigation of distributed learning for automated lesion detection in head MR images.” Radiological Physics and Technology. 2024.
41. Yamada A, Hanaoka S, Takenaga T, Yoshikawa T and Nomura Y. “Preliminary study of maximizing performance for each site in training of computer-aided diagnosis using federated learning.” CARS 2024 the 38th International Congress and Exhibition of Computer Assisted Radiology and Surgery, Barcelona, Spain, June 18–21, 2024 (Int J Comput Assist Radio Surg (proc. CARS2024) 19(suppl.1): S91-S92). 2024.
42. Xiaohang Xu, Renhe Jiang, Chuang Yang, Zipei Fan and Kaoru Sezaki. “Taming the Long Tail in Human Mobility Prediction.” 38th Annual Conference on Neural Information Processing Systems (NeurIPS). 2024.
43. Shouhei Hanaoka, Toshihiro Hanawa, Tomomi Takenaga and Yukihiro Nomura. “Secure AI model training for lung nodule detection on confidential virtual machine including GPU.” Computational Approaches for Cancer Workshop (CAFCW24) in conjunction with SC24, Poster. 2024.
44. Yamada A, Hanaoka S, Takenaga T, Yoshikawa T, Nakaguchi T and Nomura Y. “Maximizing performance for each site in automated detection of brain metastasis on contrast-enhanced T1-weight MRI using federated learning.” _ CARS 2025 the 39th International Congress and Exhibition of Computer Assisted Radiology and Surgery, Berlin, Germany, June 17–20, 2025 (Int J Comput Assist Radio Surg (proc. CARS2025) 20(suppl.1): S52-S53)._ 2025.
45. Chuang Yang, Renhe Jiang, Xiaohang Xu, Chuan Xiao and Kaoru Sezaki. “SIMformer: Single-Layer Vanilla Transformer Can Learn Free-Space Trajectory Similarity.” 51st International Conference on Very Large Data Bases (VLDB). 2025.
46. 山田 藍樹, 花岡 昇平, 竹永 智美, 吉川 健啓, 阿部 修, 野村 行弘 and 中口 俊哉. “コンピュータ支援検出／診断の学習に向けた連合学習下でのハイパーパラメータ探索の初期検討.” 第3回 JAMIT若手医用画像工学シンポジウム SAMIT2025, P2-1, (2025.11.30, 千葉). 2025.
47. Haotian Gao, Zheng Dong, Jiawei Yong, Shintaro Fukushima, Kenjiro Taura and Renhe Jiang. “How Different from the Past? Spatio-Temporal Time Series Forecasting with Self-Supervised Deviation Learning.” 39th Annual Conference on Neural Information Processing Systems (NeurIPS). 2025.
48. Aiki Yamada, Shouhei Hanaoka, Tomomi Takenaga, Soichiro Miki, Takeharu Yoshikawa, Osamu Abe, Toshiya Nakaguchi and Yukihiro Nomura. “Improving personalized federated learning to optimize site-specific performance in computer-aided detection/diagnosis.” Journal of Medical Imaging. 2025.
49. Shang Liu, Hao Du, Yang Cao, Bo Yan, Jinfei Liu and Masatoshi Yoshikawa. “PGB: Benchmarking Differentially Private Synthetic Graph Generation Algorithms.” _ International Conference on Data Engineering ._ 2025.
50. Hiroyuki Yoshida, Yuichi Sugiyama and Ryota Shioya. “AceCov: Auxiliary Composite Edge Coverage for Fuzzing.” IEEE European Symposium on Security and Privacy (EuroS&P). 2025.
51. Chang Su and Toshihiro Hanawa. “Secure Storage with Ciphertext-Policy Attribute-based Access Control for HPC System.” 情報処理学会研究報告 (2025-HPC-200) No. 21, p.1-11. 2025.
52. 椎名 峻平, 中谷 翔, 平岡 拓海 and 田浦 健次朗. “PrivJail: 差分プライバシーを強制するPythonライブラリ.” DEIM2025 第17回データ工学と情報マネジメントに関するフォーラム. 2025.
53. Shumpei Shiina, Sho Nakatani, Takumi Hiraoka and Kenjiro Taura. “PrivJail: Enforcing Differential Privacy in Pythonic Data Processing.” TPDP 2025 - Theory and Practice of Differential Privacy（ポスター）. 2025.
54. 椎名 峻平, 中谷 翔, 平岡 拓海 and 田浦 健次朗. “egRPC: Python関数呼び出しを容易にgRPC化するライブラリの設計とその応用.” xSIG 2025（ポスター）. 2025.
55. 椎名 峻平 and 田浦 健次朗. “Pythonデータフレーム演算におけるユーザレベル差分プライバシーの保証.” コンピュータセキュリティシンポジウム2025（ポスター）. 2025.
56. Wen Xu, Pengpeng Qiao, Shang Liu, Zhirun Zheng, Yang Cao and Zhetao Li. “Continuous Publication of Weighted Graphs with Local Differential Privacy.” Proceedings of the VLDB Endowment. 2025.
57. Hao Du, Shang Liu, Lele Zheng, Yang Cao, A. Nakamura and Lei Chen. “Privacy in Fine-Tuning Large Language Models: Attacks, Defenses, and Future Directions.” Lecture notes in computer science. 2025.
58. Lele Zheng, Yang Cao, Masatoshi Yoshikawa, Yulong Shen, Essam A. Rashed, Kenjiro Taura, Shouhei Hanaoka and Tao Zhang. “Sensitivity-Aware Differential Privacy for Federated Medical Imaging.” Sensors. 2025.
59. Bo Yan, Shan He, Yang Cheng, Shang Liu, Yang Cao and Chuan Shi. “Federated Graph Condensation with Information Bottleneck Principles.” Proceedings of the AAAI Conference on Artificial Intelligence. 2025.
60. Lele Zheng, Yang Cao, Leo Yu Zhang, Wei Wang, Yulong Shen and Xiaochun Cao. “MMGIA: Gradient Inversion Attack Against Multimodal Federated Learning via Intermodal Correlation.” International Joint Conferences on Artificial Intelligence (IJCAI). 2025.
61. Aiki Yamada, Shouhei Hanaoka, Tomomi Takenaga, Soichiro Miki, Takeharu Yoshikawa, Osamu Abe, Toshiya Nakaguchi and Yukihiro Nomura. “Improving personalized federated learning to optimize site-specific performance in computer-aided detection/diagnosis.” Journal of medical imaging. 2025.
62. 花岡 昇平. “医用画像異常検知のためのオートエンコーダにおける新しい情報量ボトルネックレイヤーの設計の試み.” 第4回 日本医用画像電子情報・人工知能研究会. 2025.
63. Takumi Hiraoka, Shumpei Shiina and Kenjiro Taura. “A Programming Framework for Estimating Privacy Loss in Differential Privacy.” TPDP 2025 - Theory and Practice of Differential Privacy（ポスター）. 2025.
64. Nomura Y, Hanaoka S, Yamada A, Takenaga T, Nakao T, Nakaguchi T, Yoshikawa T and Abe O. “Performance changes in federated learning for cerebral aneurysm detection in head MR angiography images with increasing numbers of institutions.” International Forum on Medical Imaging in Asia 2026 (IFMIA 2026), Jan 13-14, Kaohsiung, Taiwan (accepted). 2026.
65. Yuan Zhao, Hualei Zhu, Tingyu Jiang, Shen Li, Xiaohang Xu and Hao Henry Wang. “Co-EPG: A Framework for Co-Evolution of Planning and Grounding in Autonomous GUI Agents.” The 40th Annual AAAI Conference on Artificial Intelligence. 2026.
66. Tingyu Jiang, Shen Li, Yiyao Song, Lan Zhang, Hualei Zhu, Yuan Zhao, Xiaohang Xu, Kenjiro Taura and Hao Henry Wang. “Importance-Aware Data Selection for Efficient LLM Instruction Tuning.” The 40th Annual AAAI Conference on Artificial Intelligence (Oral). 2026.
67. 椎名 峻平, 中谷 翔, 平岡 拓海 and 田浦 健次朗. “PrivJail: 差分プライバシーを強制するPythonライブラリ.” 日本データベース学会和文論文誌 Vol. 24-J No. 3. 2026.
68. Xianjie Wu, Xiaohang Xu, Tingyu Jiang, Jian Yang, Di Liang, Xianfu Cheng, Zhenhe Wu, Linzheng Chai, Wei Zhang, Jiaheng Liu, Ge Zhang, Bob Simons, Tongliang Li and Zhoujun Li. “MMTableBench: A Multi-level Multimodal Benchmark for Reasoning and Layout Complexity in Table QA.” Proceedings of the ACM Web Conference 2026 (oral). 2026.
69. Yusong Wang, Chuang Yang, Jiawei Wang, Xiaohang Xu, Jiayi Xu, Dongyuan Li, Chuan Xiao and Renhe Jiang. “ELLMob: Event-Driven Human Mobility Generation with Self-Aligned LLM Framework.” Proceedings of the Fourteenth International Conference on Learning Representations. 2026.
70. Yunze Xiao, Tingyu He, Lionel Z. Wang, Yiming Ma, Xingyu Song, Xiaohang Xu, Mona T. Diab, Irene Li and Ka Chung Ng. “JiraiBench: A Bilingual Benchmark for Evaluating Large Language Models’ Detection of Human risky health behavior Content in Jirai Community.” Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (oral). 2026.
71. Md. Ashraful Alam, Shouhei Hanaoka, Tomomi Takenaga, Takeharu Yoshikawa and Osamu Abe. “Estimation of future occurrence of hemoglobin-A1c elevation with and without differential privacy.” BMC Medical Informatics and Decision Making. 2026.
72. 村木乃乃香, 杉山優一 and 塩谷亮太. “制御フローグラフの探索具合に基づきコーパスへの保存を制御するファジング.” 電子情報通信学会 技術研究報告 vol. 125, no. 401, CPSY2025-74, pp. 131-136. 2026.