Faculty Profiles - CAO YANG

写真a

CAO YANG

Organization

School of Computing Associate Professor

Homepage

Profile

Yang Cao is an Associate Professor at the Department of Computer Science, Tokyo Institute of Technology (Tokyo Tech), and directing the Trustworthy Data Science Lab. He is passionate about studying and teaching on algorithmic trustworthiness in data science and AI. Two of his papers on data privacy were selected as best paper finalists in top-tier conferences IEEE ICDE 2017 and ICME 2020. He was a recipient of the IEEE Computer Society Japan Chapter Young Author Award 2019, Database Society of Japan Kambayashi Young Researcher Award 2021. His research projects were/are supported by JSPS, JST, MSRA, KDDI, LINE, WeBank, etc.

https://scholar.google.com/citations?hl=en&user=S-p4DFMAAAAJJ

https://cao-lab.org

External link

Degree

Ph.D. ( 2017 Kyoto University )

Research Interests

Federated Learning
Differential Privacy
Trustworthy AI
Data Market
Privacy-Preserving Machine Learning
privacy-preserving AI
Blockchain

Research Areas

Informatics / Intelligent informatics
Informatics / Database
Informatics / Information security

Research History

Institute of Science Tokyo School of Computing

2024.10

　 More details

researchmap

Committee Memberships

International Conference Advanced Data Mining and Applications (ADMA) 2025 Program Chair

2025.10

　 More details

researchmap
The Thirteenth International Conference on Advanced Cloud and Big Data (CBD) 2025 General Chair

2025.10

　 More details

researchmap
Springer World Wide Web Journal (WWWJ) Associate Editor

2025.8

　 More details

researchmap
ACM Conference on Computer and Communications Security (CCS) Program Committee Member

2025.1 - 2025.12

　 More details

researchmap
Springer Data Science and Engineering (DSEJ) Associate Editor

2024.10

　 More details

researchmap
IEEE Transactions on Dependable and Secure Computing (TDSC) Associate Editor

2024.8 - 2026.8

　 More details

researchmap
IPSJ Transactions on Data Processing (IPSJ-TOD) Associate Editor

2024.5

　 More details

researchmap
ACM SIGMOD International Conference on Management of Data (SIGMOD) Program Committee Member

2024.1

　 More details

researchmap
ACM SIGKDDD Conference on Knowledge Discovery and Data Mining (KDD) Program Committee Member

2024.1

　 More details

researchmap
IEEE Open Journal of the Computer Society (OJCS) Associate Editor

2024.1 - 2025.12

　 More details

researchmap
International Conference on Very Large Data Bases (VLDB) Program Committee Member

2023.1

　 More details

researchmap

▼display all

Papers

CODE ACROSTIC: Robust Watermarking for Code Generation Reviewed International coauthorship International journal

Li Lin, Siyuan Xin, Yang Cao, Xiaochun Cao

Web Information Systems Engineering – WISE 2025 521 - 531 2025.12

　More details

Language：English Publishing type：Research paper (international conference proceedings) Publisher：Springer Nature Singapore

DOI： 10.1007/978-981-95-7251-9_35

researchmap
Byzantine-Robust Federated Learning Framework via a Server-Client Defense Mechanisms. Reviewed International coauthorship International journal

Anxiao Song, Tao Zhang 0029, Ke Cheng 0001, Yang Cao 0011, Xinghui Zhu, Yulong Shen 0001

IEEE Internet Things J. 12 ( 14 ) 29073 - 29088 2025.7

　More details

Language：English Publishing type：Research paper (scientific journal)

DOI： 10.1109/JIOT.2025.3568443

researchmap
Continuous Publication of Weighted Graphs with Local Differential Privacy. Reviewed International coauthorship International journal

Wen Xu, Pengpeng Qiao, Shang Liu 0001, Zhirun Zheng, Yang Cao 0011, Zhetao Li

Proceedings of the VLDB Endowment 18 ( 11 ) 4214 - 4226 2025.7

　More details

Language：English Publishing type：Research paper (scientific journal)

DOI： 10.14778/3749646.3749688

researchmap

Other Link： https://dblp.org/db/journals/pvldb/pvldb18.html#XuQLZCL25
Privacy in Fine-Tuning Large Language Models: Attacks, Defenses, and Future Directions. Reviewed International coauthorship International journal

Hao Du, Shang Liu 0001, Lele Zheng, Yang Cao 0011, Atsuyoshi Nakamura, Lei Chen

Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD) 326 - 344 2025.6

　More details

Authorship：Corresponding author Language：English Publishing type：Research paper (international conference proceedings) Publisher：Springer

DOI： 10.1007/978-981-96-8183-9_25

researchmap

Other Link： https://dblp.uni-trier.de/rec/conf/pakdd/2025-4
Sensitivity-Aware Differential Privacy for Federated Medical Imaging Reviewed International coauthorship International journal

Lele Zheng, Yang Cao, Masatoshi Yoshikawa, Yulong Shen, Essam A. Rashed, Kenjiro Taura, Shouhei Hanaoka, Tao Zhang

Sensors 25 ( 9 ) 2847 - 2847 2025.4

　More details

Language：English Publishing type：Research paper (scientific journal) Publisher：Mdpi Ag

Federated learning (FL) enables collaborative model training across multiple institutions without the sharing of raw patient data, making it particularly suitable for smart healthcare applications. However, recent studies revealed that merely sharing gradients provides a false sense of security, as private information can still be inferred through gradient inversion attacks (GIAs). While differential privacy (DP) provides provable privacy guarantees, traditional DP methods apply uniform protection, leading to excessive protection for low-sensitivity data and insufficient protection for high-sensitivity data, which degrades model performance and increases privacy risks. This paper proposes a new privacy notion, sensitivity-aware differential privacy, to better balance model performance and privacy protection. Our idea is that the sensitivity of each data sample can be objectively measured using real-world attacks. To implement this new notion, we develop the corresponding defense mechanism that adjusts privacy protection levels based on the variation in the privacy leakage risks of gradient inversion attacks. Furthermore, the method extends naturally to multi-attack scenarios. Extensive experiments on real-world medical imaging datasets demonstrate that, under equivalent privacy risk, our method achieves an average performance improvement of 13.5% over state-of-the-art methods.

DOI： 10.3390/s25092847

researchmap
Anti-Fake Vaccine: Safeguarding Privacy Against Face Swapping via Visual-Semantic Dual Degradation. Reviewed International coauthorship International journal

Jingzhi Li 0002, Changjiang Luo, Hua Zhang 0008, Yang Cao 0011, Xin Liao, Xiaochun Cao

Int. J. Comput. Vis. 133 ( 4 ) 2025 - 2043 2025.4

　More details

Language：English Publishing type：Research paper (scientific journal)

DOI： 10.1007/s11263-024-02259-5

researchmap
DPAF: Image Synthesis via Differentially Private Aggregation in Forward Phase. Reviewed International coauthorship International journal

Chih-Hsun Lin, Chia-Yi Hsu, Chia-Mu Yu, Yang Cao 0011, Chun-Ying Huang

IEEE Internet Things J. 12 ( 6 ) 7549 - 7563 2025.3

　More details

Language：English Publishing type：Research paper (scientific journal)

DOI： 10.1109/JIOT.2024.3496920

researchmap
Are Your LLM-based Text-to-SQL Models Secure? Exploring SQL Injection via Backdoor Attacks.

Meiyu Lin, Haichuan Zhang, Jiale Lao, Renyuan Li, Yuanchun Zhou, Carl Yang 0001, Yang Cao 0011, Mingjie Tang

Proceedings of the ACM on Management of Data 3 ( 6 ) 1 - 27 2025

　More details

Publishing type：Research paper (scientific journal)

DOI： 10.1145/3769762

researchmap
Differentially Private Visual Learning with Public Subspace Augmented by Synthetic Data. Reviewed International coauthorship International journal

Haichao Sha, Yuncheng Wu, Ruixuan Liu, Yang Cao 0011, Hong Chen 0001

ACM Multimedia 11357 - 11366 2025

　More details

Language：English Publishing type：Research paper (international conference proceedings)

DOI： 10.1145/3746027.3755037

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/mm/mm2025.html#ShaWL0025
PGB: Benchmarking Differentially Private Synthetic Graph Generation Algorithms. Reviewed International coauthorship International journal

Shang Liu 0001, Hao Du, Yang Cao 0011, Bo Yan 0005, Jinfei Liu, Masatoshi Yoshikawa

International Conference on Data Engineering (ICDE) 1348 - 1361 2025

　More details

Language：English Publishing type：Research paper (international conference proceedings) Publisher：Ieee

DOI： 10.1109/ICDE65448.2025.00105

researchmap

Other Link： https://dblp.uni-trier.de/rec/conf/icde/2025
Bargaining-Based Data Markets Reviewed International coauthorship International journal

Yuran Bi, Jinfei Liu, Kui Ren 0001, Yihang Wu, Yang Cao 0011

International Conference on Data Engineering (ICDE) 1456 - 1469 2025

　More details

Language：English Publishing type：Research paper (international conference proceedings) Publisher：Ieee

DOI： 10.1109/ICDE65448.2025.00113

researchmap

Other Link： https://dblp.uni-trier.de/rec/conf/icde/2025
Enhancing Federated Domain Adaptation with Multi-Domain Prototype-Based Federated Fine-Tuning. Reviewed International coauthorship International journal

Jingyuan Zhang, Yiyang Duan, Shuaicheng Niu, Yang Cao 0011, Wei Yang Bryan Lim

International Conference on Learning Representations (ICLR) 2025

　More details

Language：English Publishing type：Research paper (international conference proceedings) Publisher：OpenReview.net

researchmap

Other Link： https://openreview.net/forum?id=3wEGdrV5Cb
Federated Graph Condensation with Information Bottleneck Principles. Reviewed International coauthorship International journal

Bo Yan 0005, Sihao He, Cheng Yang 0002, Shang Liu 0001, Yang Cao 0011, Chuan Shi 0001

AAAI 12990 - 12998 2025

　More details

Language：English Publishing type：Research paper (international conference proceedings) Publisher：AAAI Press

DOI： 10.1609/aaai.v39i12.33417

researchmap

Other Link： https://dblp.uni-trier.de/rec/conf/aaai/2025
MMGIA: Gradient Inversion Attack Against Multimodal Federated Learning via Intermodal Correlation. Reviewed International coauthorship International journal

Lele Zheng, Yang Cao 0011, Leo Yu Zhang, Wei Wang 0077, Yulong Shen 0001, Xiaochun Cao

Ijcai 7967 - 7975 2025

　More details

Language：English Publishing type：Research paper (international conference proceedings) Publisher：Ijcai.org

DOI： 10.24963/ijcai.2025/886

researchmap

Other Link： https://dblp.uni-trier.de/rec/conf/ijcai/2025
Pursuit of Truth: Incentive Mechanism Involving Privacy Demands in Mobile Crowdsourcing. International coauthorship International journal

Ping Wang 0045, Pengpeng Qiao, Yang Cao 0011, Tao Feng 0007

Adc 94 - 107 2025

　More details

Language：English Publishing type：Research paper (international conference proceedings)

DOI： 10.1007/978-981-95-6196-4_7

researchmap

Other Link： https://dblp.org/db/conf/adc/adc2025.html#WangQCF25
AegisGuard: RL-Guided Adapter Tuning for TEE-Based Efficient & Secure On-Device Inference Reviewed International coauthorship

CHE WANG, Ziqi Zhang, Yinggui Wang, Tiantong Wang, Yurong Hao, Jianbo Gao, Tao Wei, YANG CAO, Zhong Chen, Wei Yang, Bryan Lim

NeurIPS 2025

　More details

Language：English Publishing type：Research paper (international conference proceedings)

researchmap

▼display all

Awards

Best Paper Award

2025.12 Australasian Database Conference (ADC) "Pursuit of Truth: Incentive Mechanism Involving Privacy Demands in Mobile Crowdsourcing"

Ping Wang, Pengpeng Qiao, Yang Cao, Tao Feng

　More details

Award type：Award from international society, conference, symposium, etc.

researchmap
Outstanding Paper Award

2025.10 ACM International Conference on Multimedia "Differentially Private Visual Learning with Public Subspace Augmented by Synthetic Data"

Haichao Sha, Yuncheng Wu, Ruixuan Liu, YANG CAO, Hong Chen

　More details

Award type：Award from international society, conference, symposium, etc.

researchmap

Research Projects

大規模言語モデルのための新しい信頼性向上技術

2023 - 2026

科学技術振興機構戦略的な研究開発の推進戦略的創造研究推進事業さきがけ

曹洋

　 More details

Authorship：Principal investigator

大規模言語モデル(LLMs)は私たちの社会を変革していますが、プライバシーや堅牢性、悪用のリスクももたらしています。本研究は、大規模言語モデルの信頼性を向上させる新しい技術を開発することを目指しており、敵対的な環境での大規模言語モデルの堅牢性を向上させ、大規模言語モデルによって生成されたコンテンツの誤用を防ぐことを目的としています。

researchmap

J-GLOBAL
A Principled Framework for Explaining, Choosing and Negotiating Privacy Parameters of Differential Privacy

Grant number：23K24851 2022.4 - 2025.3

JSPS Grants-in-Aid for Scientific Research (B) Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B)

　 More details

Authorship：Principal investigator

Grant amount：\17160000 （ Direct Cost: \13200000 、 Indirect Cost：\3960000 ）

researchmap
分散型ソーシャルグラフに向けた差分プライバシー技術 (分担)

Grant number：22H00521 2022.3 - 2027.4

Japan Society for the Promotion of Science Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (A)

　 More details

Authorship：Coinvestigator(s)

Grant amount：\41860000 （ Direct Cost: \32200000 、 Indirect Cost：\9660000 ）

researchmap
統計的に異質なエッジクラウドネットワークにおける効率的かつプライバシー保護された大規模マルチモーダルモデルの学習と推論

Grant number：25126435 2025 - 2027

科学技術振興機構国際的な科学技術共同研究などの推進/日ASEAN科学技術・イノベーション協働連携事業（NEXUS）/国際共同研究

曹洋, 蘇春華, RASHED Essam

　 More details

本研究は、エッジクラウドネットワークにおいて、大規模マルチモーダルモデル（ＬＭＭ）の効率的かつプライバシー保護された学習と推論の実現を目的とする。
具体的には、日本側チームは、連合学習によりプライバシー保護技術を開発し、特にメトリック差分プライバシーや信頼実行環境を使い、ＬＭＭの安全性向上を担い、シンガポール側チームは、ＬＭＭの効率的な推論のため、エッジとクラウド間データスクを動的に割り振るＭｉｘｔｕｒｅｏｆＥｘｐｅｒｔｓ（ＭｏＥ）やＲｅｔｒｉｅｖａｌ－ＡｕｇｍｅｎｔｅｄＧｅｎｅｒａｔｉｏｎ（ＲＡＧ）のフレームワーク構築を実施する。
両国の研究チームによる共同研究を通じて、高性能かつ低遅延のＡＩシステムが開発され、国際的な技術進展に貢献することを目指す。

researchmap