Masatoshi Yoshikawa
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Masatoshi Yoshikawa has a total of 125 co-authors. Here are the top co-authors by number of shared publications:
- Yang Cao
Tokyo Institute of Technology
26 shared publications - Li Xiong
Emory University
7 shared publications Yang Cao
Kyoto University
7 shared publicationsYasuhito Asano
Department of Social Informatics, Kyoto University, Kyoto, Japan
7 shared publications- Adam Jatowt
University of Innsbruck
6 shared publications - Fumiyuki Kato
Kyoto University Emory University Kyoto University Hokkaido University Osaka Seikei University
5 shared publications
Latest Publications
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 23 (6) , 11857-11871
Federated Graph Analytics With Differential Privacy
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 23 (6) , 11857-11871
Federated Graph Analytics With Differential Privacy
- Open Access
- 30 July 2026
- Published by Scilight Press Pty Ltd in Transactions on Graph Intelligence and Network Applications
Differentially Private Graph Learning: A Survey
Graph learning has been widely applied across diverse domains, including social networks, recommendation systems, and bioinformatics. However, real-world graph data often contains highly sensitive information, raising serious privacy concerns. Differential Privacy (DP) has emerged as a rigorous mathematical framework to protect sensitive information in graph learning while providing formal privacy guarantees. This survey presents the first comprehensive and systematic review of Differentially Private Graph Learning (DPGL). We organize existing DPGL methods into four categories based on the granularity of privacy protection, namely node-level, edge-level, graph-level, and local differential privacy. Within each category, we further analyze learning paradigms, perturbation mechanisms, and their respective strengths and limitations, and identify key technical challenges in the field. Furthermore, we identify future research directions critical for advancing DPGL toward practical deployment in real-world applications. This survey aims to provide a unified reference for researchers and practitioners while inspiring future innovations in privacy-preserving graph learning.
- 30 July 2026
- Published by Scilight Press Pty Ltd in Transactions on Graph Intelligence and Network Applications
Open AccessDifferentially Private Graph Learning: A Survey
Graph learning has been widely applied across diverse domains, including social networks, recommendation systems, and bioinformatics. However, real-world graph data often contains highly sensitive information, raising serious privacy concerns. Differential Privacy (DP) has emerged as a rigorous mathematical framework to protect sensitive information in graph learning while providing formal privacy guarantees. This survey presents the first comprehensive and systematic review of Differentially Private Graph Learning (DPGL). We organize existing DPGL methods into four categories based on the granularity of privacy protection, namely node-level, edge-level, graph-level, and local differential privacy. Within each category, we further analyze learning paradigms, perturbation mechanisms, and their respective strengths and limitations, and identify key technical challenges in the field. Furthermore, we identify future research directions critical for advancing DPGL toward practical deployment in real-world applications. This survey aims to provide a unified reference for researchers and practitioners while inspiring future innovations in privacy-preserving graph learning.
- Open Access
Sensitivity-Aware Differential Privacy for Federated Medical Imaging
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.
Open AccessSensitivity-Aware Differential Privacy for Federated Medical Imaging
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.
- Open Access
Extracting Spatiotemporal Data from Gradients with Large Language Models
Recent works show that sensitive user data can be reconstructed from gradient updates, breaking the key privacy promise of federated learning. While success was demonstrated primarily on image data, these methods do not directly transfer to other domains, such as spatiotemporal data. To understand privacy risks in spatiotemporal federated learning, we first propose Spatiotemporal Gradient Inversion Attack (ST-GIA), a gradient attack algorithm tailored to spatiotemporal data that successfully reconstructs the original location from gradients. Furthermore, the absence of priors in attacks on spatiotemporal data has hindered the accurate reconstruction of real client data. To address this limitation, we propose ST-GIA+, which utilizes an auxiliary language model to guide the search for potential locations, thereby successfully reconstructing the original data from gradients. In addition, we design an adaptive defense strategy to mitigate gradient inversion attacks in spatiotemporal federated learning. By dynamically adjusting the perturbation levels, we can offer tailored protection for varying rounds of training data, thereby achieving a better trade-off between privacy and utility than current state-of-the-art methods. Through intensive experimental analysis on three real-world datasets, we reveal that the proposed defense strategy can well preserve the utility of spatiotemporal federated learning with effective security protection.
Open AccessExtracting Spatiotemporal Data from Gradients with Large Language Models
Recent works show that sensitive user data can be reconstructed from gradient updates, breaking the key privacy promise of federated learning. While success was demonstrated primarily on image data, these methods do not directly transfer to other domains, such as spatiotemporal data. To understand privacy risks in spatiotemporal federated learning, we first propose Spatiotemporal Gradient Inversion Attack (ST-GIA), a gradient attack algorithm tailored to spatiotemporal data that successfully reconstructs the original location from gradients. Furthermore, the absence of priors in attacks on spatiotemporal data has hindered the accurate reconstruction of real client data. To address this limitation, we propose ST-GIA+, which utilizes an auxiliary language model to guide the search for potential locations, thereby successfully reconstructing the original data from gradients. In addition, we design an adaptive defense strategy to mitigate gradient inversion attacks in spatiotemporal federated learning. By dynamically adjusting the perturbation levels, we can offer tailored protection for varying rounds of training data, thereby achieving a better trade-off between privacy and utility than current state-of-the-art methods. Through intensive experimental analysis on three real-world datasets, we reveal that the proposed defense strategy can well preserve the utility of spatiotemporal federated learning with effective security protection.
- 1 October 2024
- Published by Springer Nature
- p. 457-473
Enhancing Privacy of Spatiotemporal Federated Learning Against Gradient Inversion Attacks
- 1 October 2024
- Published by Springer Nature
- p. 457-473
Enhancing Privacy of Spatiotemporal Federated Learning Against Gradient Inversion Attacks
- Open Access
CARGO: Crypto-Assisted Differentially Private Triangle Counting without Trusted Servers
Differentially private triangle counting in graphs is essential for analyzing connection patterns and calculating clustering coefficients while protecting sensitive individual information. Previous works have relied on either central or local models to enforce differential privacy. However, a significant utility gap exists between the central and local models of differentially private triangle counting, depending on whether or not a trusted server is needed. In particular, the central model provides a high accuracy but necessitates a trusted server. The local model does not require a trusted server but suffers from limited accuracy. Our paper introduces a crypto-assisted differentially private triangle counting system, named CARGO, leveraging cryptographic building blocks to improve the effectiveness of differentially private triangle counting without assumption of trusted servers. It achieves high utility similar to the central model but without the need for a trusted server like the local model. CARGO consists of three main components. First, we introduce a similarity-based projection method that reduces the global sensitivity while preserving more triangles via triangle homogeneity. Second, we present a triangle counting scheme based on the additive secret sharing that securely and accurately computes the triangles while protecting sensitive information. Third, we design a distributed perturbation algorithm that perturbs the triangle count with minimal but sufficient noise. We also provide a comprehensive theoretical and empirical analysis of our proposed methods. Extensive experiments demonstrate that our CARGO significantly outperforms the local model in terms of utility and achieves high-utility triangle counting comparable to the central model.
Open AccessCARGO: Crypto-Assisted Differentially Private Triangle Counting without Trusted Servers
Differentially private triangle counting in graphs is essential for analyzing connection patterns and calculating clustering coefficients while protecting sensitive individual information. Previous works have relied on either central or local models to enforce differential privacy. However, a significant utility gap exists between the central and local models of differentially private triangle counting, depending on whether or not a trusted server is needed. In particular, the central model provides a high accuracy but necessitates a trusted server. The local model does not require a trusted server but suffers from limited accuracy. Our paper introduces a crypto-assisted differentially private triangle counting system, named CARGO, leveraging cryptographic building blocks to improve the effectiveness of differentially private triangle counting without assumption of trusted servers. It achieves high utility similar to the central model but without the need for a trusted server like the local model. CARGO consists of three main components. First, we introduce a similarity-based projection method that reduces the global sensitivity while preserving more triangles via triangle homogeneity. Second, we present a triangle counting scheme based on the additive secret sharing that securely and accurately computes the triangles while protecting sensitive information. Third, we design a distributed perturbation algorithm that perturbs the triangle count with minimal but sufficient noise. We also provide a comprehensive theoretical and empirical analysis of our proposed methods. Extensive experiments demonstrate that our CARGO significantly outperforms the local model in terms of utility and achieves high-utility triangle counting comparable to the central model.
- Open Access
- 8 September 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (18) , 10132
Local Differential Privacy Image Generation Using Flow-Based Deep Generative Models
Diagnostic radiologists need artificial intelligence (AI) for medical imaging, but access to medical images required for training in AI has become increasingly restrictive. To release and use medical images, we need an algorithm that can simultaneously protect privacy and preserve pathologies in medical images. To address this, we introduce DP-GLOW, a hybrid that combines the local differential privacy (LDP) algorithm with GLOW, one of the flow-based deep generative models. By applying a GLOW model, we disentangle the pixelwise correlation of images, which makes it difficult to protect privacy with straightforward LDP algorithms for images. Specifically, we map images to the latent vector of the GLOW model, where each element follows an independent normal distribution. We then apply the Laplace mechanism to this latent vector to achieve -LDP, which is one of the LDP algorithms. Moreover, we applied DP-GLOW to chest X-ray images to generate LDP images while preserving pathologies. The -LDP-processed chest X-ray images obtained with DP-GLOW indicate that we have obtained a powerful tool for releasing and using medical images for training AI.- 8 September 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (18) , 10132
Open AccessLocal Differential Privacy Image Generation Using Flow-Based Deep Generative Models
Diagnostic radiologists need artificial intelligence (AI) for medical imaging, but access to medical images required for training in AI has become increasingly restrictive. To release and use medical images, we need an algorithm that can simultaneously protect privacy and preserve pathologies in medical images. To address this, we introduce DP-GLOW, a hybrid that combines the local differential privacy (LDP) algorithm with GLOW, one of the flow-based deep generative models. By applying a GLOW model, we disentangle the pixelwise correlation of images, which makes it difficult to protect privacy with straightforward LDP algorithms for images. Specifically, we map images to the latent vector of the GLOW model, where each element follows an independent normal distribution. We then apply the Laplace mechanism to this latent vector to achieve -LDP, which is one of the LDP algorithms. Moreover, we applied DP-GLOW to chest X-ray images to generate LDP images while preserving pathologies. The -LDP-processed chest X-ray images obtained with DP-GLOW indicate that we have obtained a powerful tool for releasing and using medical images for training AI. - Open Access
- 28 July 2023
- Published by Springer Nature in International Journal on Digital Libraries
- Vol. 25 (2) , 197-216
Is this news article still relevant? Ranking by contemporary relevance in archival search
Our civilization creates enormous volumes of digital data, a substantial fraction of which is preserved and made publicly available for present and future usage. Additionally, historical born-analog records are progressively being digitized and incorporated into digital document repositories. While professionals often have a clear idea of what they are looking for in document archives, average users are likely to have no precise search needs when accessing available archives (e.g., through their online interfaces). Thus, if the results are to be relevant and appealing to average people, they should include engaging and recognizable material. However, state-of-the-art document archival retrieval systems essentially use the same approaches as search engines for synchronic document collections. In this article, we develop unique ranking criteria for assessing the usefulness of archived contents based on their estimated relationship with current times, which we call contemporary relevance. Contemporary relevance may be utilized to enhance access to archival document collections, increasing the likelihood that users will discover interesting or valuable material. We next present an effective strategy for estimating contemporary relevance degrees of news articles by utilizing learning to rank approach based on a variety of diverse features, and we then successfully test it on the New York Times news collection. The incorporation of the contemporary relevance computation into archival retrieval systems should enable a new search style in which search results are meant to relate to the context of searchers’ times, and by this have the potential to engage the archive users. As a proof of concept, we develop and demonstrate a working prototype of a simplified ranking model that operates on the top of the Portuguese Web Archive portal (arquivo.pt).
- 28 July 2023
- Published by Springer Nature in International Journal on Digital Libraries
- Vol. 25 (2) , 197-216
Open AccessIs this news article still relevant? Ranking by contemporary relevance in archival search
Our civilization creates enormous volumes of digital data, a substantial fraction of which is preserved and made publicly available for present and future usage. Additionally, historical born-analog records are progressively being digitized and incorporated into digital document repositories. While professionals often have a clear idea of what they are looking for in document archives, average users are likely to have no precise search needs when accessing available archives (e.g., through their online interfaces). Thus, if the results are to be relevant and appealing to average people, they should include engaging and recognizable material. However, state-of-the-art document archival retrieval systems essentially use the same approaches as search engines for synchronic document collections. In this article, we develop unique ranking criteria for assessing the usefulness of archived contents based on their estimated relationship with current times, which we call contemporary relevance. Contemporary relevance may be utilized to enhance access to archival document collections, increasing the likelihood that users will discover interesting or valuable material. We next present an effective strategy for estimating contemporary relevance degrees of news articles by utilizing learning to rank approach based on a variety of diverse features, and we then successfully test it on the New York Times news collection. The incorporation of the contemporary relevance computation into archival retrieval systems should enable a new search style in which search results are meant to relate to the context of searchers’ times, and by this have the potential to engage the archive users. As a proof of concept, we develop and demonstrate a working prototype of a simplified ranking model that operates on the top of the Portuguese Web Archive portal (arquivo.pt).
- Open Access
- 18 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 812-821
BiTimeBERT: Extending Pre-Trained Language Representations with Bi-Temporal Information
Time is an important aspect of documents and is used in a range of NLP and IR tasks. In this work, we investigate methods for incorporating temporal information during pre-training to further improve the performance on time-related tasks. Compared with common pre-trained language models like BERT which utilize synchronic document collections (e.g., BookCorpus and Wikipedia) as the training corpora, we use long-span temporal news article collection for building word representations. We introduce BiTimeBERT, a novel language representation model trained on a temporal collection of news articles via two new pre-training tasks, which harnesses two distinct temporal signals to construct time-aware language representations. The experimental results show that BiTimeBERT consistently outperforms BERT and other existing pre-trained models with substantial gains on different downstream NLP tasks and applications for which time is of importance (e.g., the accuracy improvement over BERT is 155% on the event time estimation task).
- 18 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 812-821
Open AccessBiTimeBERT: Extending Pre-Trained Language Representations with Bi-Temporal Information
Time is an important aspect of documents and is used in a range of NLP and IR tasks. In this work, we investigate methods for incorporating temporal information during pre-training to further improve the performance on time-related tasks. Compared with common pre-trained language models like BERT which utilize synchronic document collections (e.g., BookCorpus and Wikipedia) as the training corpora, we use long-span temporal news article collection for building word representations. We introduce BiTimeBERT, a novel language representation model trained on a temporal collection of news articles via two new pre-training tasks, which harnesses two distinct temporal signals to construct time-aware language representations. The experimental results show that BiTimeBERT consistently outperforms BERT and other existing pre-trained models with substantial gains on different downstream NLP tasks and applications for which time is of importance (e.g., the accuracy improvement over BERT is 155% on the event time estimation task).
- Open Access
OLIVE: Oblivious Federated Learning on Trusted Execution Environment against the risk of sparsification
Combining Federated Learning (FL) with a Trusted Execution Environment (TEE) is a promising approach for realizing privacy-preserving FL, which has garnered significant academic attention in recent years. Implementing the TEE on the server side enables each round of FL to proceed without exposing the client's gradient information to untrusted servers. This addresses usability gaps in existing secure aggregation schemes as well as utility gaps in differentially private FL. However, to address the issue using a TEE, the vulnerabilities of server-side TEEs need to be considered -- this has not been sufficiently investigated in the context of FL. The main technical contribution of this study is the analysis of the vulnerabilities of TEE in FL and the defense. First, we theoretically analyze the leakage of memory access patterns, revealing the risk of sparsified gradients, which are commonly used in FL to enhance communication efficiency and model accuracy. Second, we devise an inference attack to link memory access patterns to sensitive information in the training dataset. Finally, we propose an oblivious yet efficient aggregation algorithm to prevent memory access pattern leakage. Our experiments on real-world data demonstrate that the proposed method functions efficiently in practical scales.
Open AccessOLIVE: Oblivious Federated Learning on Trusted Execution Environment against the risk of sparsification
Combining Federated Learning (FL) with a Trusted Execution Environment (TEE) is a promising approach for realizing privacy-preserving FL, which has garnered significant academic attention in recent years. Implementing the TEE on the server side enables each round of FL to proceed without exposing the client's gradient information to untrusted servers. This addresses usability gaps in existing secure aggregation schemes as well as utility gaps in differentially private FL. However, to address the issue using a TEE, the vulnerabilities of server-side TEEs need to be considered -- this has not been sufficiently investigated in the context of FL. The main technical contribution of this study is the analysis of the vulnerabilities of TEE in FL and the defense. First, we theoretically analyze the leakage of memory access patterns, revealing the risk of sparsified gradients, which are commonly used in FL to enhance communication efficiency and model accuracy. Second, we devise an inference attack to link memory access patterns to sensitive information in the training dataset. Finally, we propose an oblivious yet efficient aggregation algorithm to prevent memory access pattern leakage. Our experiments on real-world data demonstrate that the proposed method functions efficiently in practical scales.