Ke Cheng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Ke Cheng has a total of 66 co-authors. Here are the top co-authors by number of shared publications:
- Yulong Shen
Xidian University
16 shared publications - Anxiao Song
School of Computer Science and Technology, Xidian University, Xian, China
8 shared publications - Tao Zhang
Xidian University
6 shared publications Jiaxuan Fu
Tencent Inc
5 shared publications- Jiaxuan Fu
Xidian University, Xi'an, China
5 shared publications - Xinghui Zhu
School of Computer Science and Technology, Xidian University, Xi’an, China
4 shared publications
Latest Publications
- 1 September 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 23 (5) , 11169-11183
SEO-DBSCAN: Securely and Efficiently Outsourcing Density-Based Clustering
- 1 September 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 23 (5) , 11169-11183
SEO-DBSCAN: Securely and Efficiently Outsourcing Density-Based Clustering
- 21 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. PP (15455971) , 1-15
FSS-DT: Secure Division-free Decision Tree Training and Inference via Function Secret Sharing
Decision tree (DT), an efficient and interpretable machine learning algorithm, has been widely used in numerous industrial applications. As the demand for higher performance grows, there is increasing interest in cross-domain collaboration for training DTs using vertical datasets, where different owners hold distinct features but share the same set of samples. Due to privacy concerns, leveraging secure multi-party computation (MPC) to train DTs jointly has gained significant attention. Although promising, deploying these approaches across large scale vertical datasets presents formidable challenges for two bottlenecks: divisions and comparisons. MPC is not friendly to process divisions, whereas DT training algorithms involve division to evaluate each feature split for each tree node. DT training and inference also involve numerous comparisons, yet existing secure comparisons are still not efficient enough. We propose FSS-DT, a highly efficient two-party framework for secure DT training and inference on vertically distributed datasets. The efficiency of FSS-DT is achieved by designing a division-free MPC-friendly DT algorithm. Moreover, we present a lightweight secure comparison protocol using our optimized distributed comparison function (DCF), an instance of function secret sharing, via “0/1-encoding” and “early termination” techniques. We have implemented a prototype of FSS-DT. Extensive experiments demonstrate that FSS-DT achieves the same accu racy as models trained in plaintext and outperforms the state-of the-art Pivot and PriVDT. Specifically, for training on LAN and WAN over the large UCI dataset, FSS-DT is 17.5× and 51.1× faster than Pivot, and about 8.4× and 8.8× faster than Swan, respectively.
- 21 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. PP (15455971) , 1-15
FSS-DT: Secure Division-free Decision Tree Training and Inference via Function Secret Sharing
Decision tree (DT), an efficient and interpretable machine learning algorithm, has been widely used in numerous industrial applications. As the demand for higher performance grows, there is increasing interest in cross-domain collaboration for training DTs using vertical datasets, where different owners hold distinct features but share the same set of samples. Due to privacy concerns, leveraging secure multi-party computation (MPC) to train DTs jointly has gained significant attention. Although promising, deploying these approaches across large scale vertical datasets presents formidable challenges for two bottlenecks: divisions and comparisons. MPC is not friendly to process divisions, whereas DT training algorithms involve division to evaluate each feature split for each tree node. DT training and inference also involve numerous comparisons, yet existing secure comparisons are still not efficient enough. We propose FSS-DT, a highly efficient two-party framework for secure DT training and inference on vertically distributed datasets. The efficiency of FSS-DT is achieved by designing a division-free MPC-friendly DT algorithm. Moreover, we present a lightweight secure comparison protocol using our optimized distributed comparison function (DCF), an instance of function secret sharing, via “0/1-encoding” and “early termination” techniques. We have implemented a prototype of FSS-DT. Extensive experiments demonstrate that FSS-DT achieves the same accu racy as models trained in plaintext and outperforms the state-of the-art Pivot and PriVDT. Specifically, for training on LAN and WAN over the large UCI dataset, FSS-DT is 17.5× and 51.1× faster than Pivot, and about 8.4× and 8.8× faster than Swan, respectively.
- Top Cited Papers
- 6 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 23 (3) , 6167-6182
PriFFT: Privacy-Preserving Federated Fine-Tuning of Large Language Models via Hybrid Secret Sharing
Fine-tuning large language models (LLMs) raises privacy concerns due to the risk of exposing sensitive training data. Federated learning (FL) mitigates this risk by keeping training samples on local devices, while facing the following privacy problems. (i) Recent studies show that adversaries can still infer private information in FL. (ii) LLM parameters are shared publicly during federated fine-tuning, while developers are often reluctant to disclose these parameters. (iii) Existing works focus on secure inference but neglects privacy-preserving fine-tuning. Inspired by above problems, we propose PriFFT, a privacy-preserving federated fine-tuning mechanism, to protect both model parameters and users' privacy. Due to considerable LLM parameters, we present hybrid secret sharing combining arithmetic secret sharing (ASS) and function secret sharing (FSS) to build secure operations and implement secure neural network layers and activation functions for fine-tuning. We optimize several secure protocols based on FSS, including reciprocal calculation, tensor products, natural exponentiation, softmax, sigmoid, hyperbolic tangent, and dropout. The hybrid secret sharing enables PriFFT to apply our optimized FSS protocols while combining ASS protocols to support complex computation without extra communication. PriFFT reduces execution time and communication overhead in privacy-preserving fine-tuning up to 59.1% and 77.0% without accuracy drop compared to the existing secret sharing methods.
- 6 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 23 (3) , 6167-6182
Top Cited PapersPriFFT: Privacy-Preserving Federated Fine-Tuning of Large Language Models via Hybrid Secret Sharing
Fine-tuning large language models (LLMs) raises privacy concerns due to the risk of exposing sensitive training data. Federated learning (FL) mitigates this risk by keeping training samples on local devices, while facing the following privacy problems. (i) Recent studies show that adversaries can still infer private information in FL. (ii) LLM parameters are shared publicly during federated fine-tuning, while developers are often reluctant to disclose these parameters. (iii) Existing works focus on secure inference but neglects privacy-preserving fine-tuning. Inspired by above problems, we propose PriFFT, a privacy-preserving federated fine-tuning mechanism, to protect both model parameters and users' privacy. Due to considerable LLM parameters, we present hybrid secret sharing combining arithmetic secret sharing (ASS) and function secret sharing (FSS) to build secure operations and implement secure neural network layers and activation functions for fine-tuning. We optimize several secure protocols based on FSS, including reciprocal calculation, tensor products, natural exponentiation, softmax, sigmoid, hyperbolic tangent, and dropout. The hybrid secret sharing enables PriFFT to apply our optimized FSS protocols while combining ASS protocols to support complex computation without extra communication. PriFFT reduces execution time and communication overhead in privacy-preserving fine-tuning up to 59.1% and 77.0% without accuracy drop compared to the existing secret sharing methods.
- Open Access
Differentially Private Subspace Fine-Tuning for Large Language Models
Fine-tuning large language models on downstream tasks is crucial for realizing their cross-domain potential but often relies on sensitive data, raising privacy concerns. Differential privacy (DP) offers rigorous privacy guarantees and has been widely adopted in fine-tuning; however, naively injecting noise across the high-dimensional parameter space creates perturbations with large norms, degrading performance and destabilizing training. To address this issue, we propose DP-SFT, a two-stage subspace fine-tuning method that substantially reduces noise magnitude while preserving formal DP guarantees. Our intuition is that, during fine-tuning, significant parameter updates lie within a low-dimensional, task-specific subspace, while other directions change minimally. Hence, we only inject DP noise into this subspace to protect privacy without perturbing irrelevant parameters. In phase one, we identify the subspace by analyzing principal gradient directions to capture task-specific update signals. In phase two, we project full gradients onto this subspace, add DP noise, and map the perturbed gradients back to the original parameter space for model updates, markedly lowering noise impact. Experiments on multiple datasets demonstrate that DP-SFT enhances accuracy and stability under rigorous DP constraints, accelerates convergence, and achieves substantial gains over DP fine-tuning baselines.
Open AccessDifferentially Private Subspace Fine-Tuning for Large Language Models
Fine-tuning large language models on downstream tasks is crucial for realizing their cross-domain potential but often relies on sensitive data, raising privacy concerns. Differential privacy (DP) offers rigorous privacy guarantees and has been widely adopted in fine-tuning; however, naively injecting noise across the high-dimensional parameter space creates perturbations with large norms, degrading performance and destabilizing training. To address this issue, we propose DP-SFT, a two-stage subspace fine-tuning method that substantially reduces noise magnitude while preserving formal DP guarantees. Our intuition is that, during fine-tuning, significant parameter updates lie within a low-dimensional, task-specific subspace, while other directions change minimally. Hence, we only inject DP noise into this subspace to protect privacy without perturbing irrelevant parameters. In phase one, we identify the subspace by analyzing principal gradient directions to capture task-specific update signals. In phase two, we project full gradients onto this subspace, add DP noise, and map the perturbed gradients back to the original parameter space for model updates, markedly lowering noise impact. Experiments on multiple datasets demonstrate that DP-SFT enhances accuracy and stability under rigorous DP constraints, accelerates convergence, and achieves substantial gains over DP fine-tuning baselines.
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 23 (5) , 1-12
Tuning Your Privacy on Outsourced Data Through Semi-Trusted Helpers
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 23 (5) , 1-12
Tuning Your Privacy on Outsourced Data Through Semi-Trusted Helpers
- Open Access
- 19 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 4124-4138
Mosformer: Maliciously Secure Three-Party Inference Framework for Large Transformers
- 19 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 4124-4138
Open AccessMosformer: Maliciously Secure Three-Party Inference Framework for Large Transformers
- 16 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 22 (6) , 6088-6103
Private Learning for Vertical Decision Trees: A Secure, Accurate, and Fast Realization
Private learning for vertical decision trees (PVDT) is an emerging paradigm that allows multiple parties to execute cooperative training and inference of decision trees on vertically partitioned datasets, without revealing either party%'s data or model. The state-of-the-art PVDT schemes employ the secret-sharing-based secure multi-party computation (MPC) to admit low computational cost and low bandwidth. Nevertheless, existing schemes need many communication rounds for computing concrete protocols in PVDT, like the less-than comparison, division, etc. This property is not suited for large-communication-latency networks such as WAN. In this work, we present a two-party PVDT framework, called Swan, to enable a secure, accurate, and fast realization of vertical decision trees. At the core of Swan, we design a secure and parallel protocol for $N$-input multiplication with one communication round. This forms the cornerstone for a series of secure and communication-efficient computation protocols specifically tailored to less-than comparison and division. Along the way, we use these optimized protocols to refine the training and inference processes of PVDT, achieving a significant reduction in both communication costs and rounds. Experimental results show Swan provides top-notch accuracy, and achieves a $10.2\times$ and $2.8\times$ improvement in online training and inference latency over WAN compared to prior art.
- 16 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 22 (6) , 6088-6103
Private Learning for Vertical Decision Trees: A Secure, Accurate, and Fast Realization
Private learning for vertical decision trees (PVDT) is an emerging paradigm that allows multiple parties to execute cooperative training and inference of decision trees on vertically partitioned datasets, without revealing either party%'s data or model. The state-of-the-art PVDT schemes employ the secret-sharing-based secure multi-party computation (MPC) to admit low computational cost and low bandwidth. Nevertheless, existing schemes need many communication rounds for computing concrete protocols in PVDT, like the less-than comparison, division, etc. This property is not suited for large-communication-latency networks such as WAN. In this work, we present a two-party PVDT framework, called Swan, to enable a secure, accurate, and fast realization of vertical decision trees. At the core of Swan, we design a secure and parallel protocol for $N$-input multiplication with one communication round. This forms the cornerstone for a series of secure and communication-efficient computation protocols specifically tailored to less-than comparison and division. Along the way, we use these optimized protocols to refine the training and inference processes of PVDT, achieving a significant reduction in both communication costs and rounds. Experimental results show Swan provides top-notch accuracy, and achieves a $10.2\times$ and $2.8\times$ improvement in online training and inference latency over WAN compared to prior art.
- 9 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12 (14) , 29073-29088
Byzantine-Robust Federated Learning Framework via a Server-Client Defense Mechanisms
Federated learning (FL), a distributed machine learning (ML) framework, is susceptible to Byzantine attacks since the attacker can manipulate clients’ local data or models to compromise the performance of the global model. There has been a wealth of defenses developed to mitigate the attacks by limiting the impact of malicious models. Nevertheless, the attacker can easily circumvent these approaches that rely solely on a single server-side defense, stemming from the high dimensionality of models and the variety of Byzantine attacks. Therefore, we propose Basalt, a Byzantine-robust FL framework with a server-client joint defense mechanism that enables multiple clients to train a global ML model under Byzantine attacks. On the client side, we design an efficient self-defense approach with model-level penalty loss that restricts local-benign divergence and decreases local-malicious correlation to prevent misclassification. On the server side, we present an efficient defense strategy based on the manifold and maximum clique, further strengthening the FL’s resilience against Byzantine attacks. We provide theoretical guarantees for global model convergence in FL with Byzantine attacks. Our extensive experiments demonstrate that Basalt outperforms existing state-of-the-art works. Especially, it achieves nearly 100% accuracy for detecting malicious clients in nonindependent and identically distributed (Non-IID) MNIST datasets under various Byzantine attacks.
- 9 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12 (14) , 29073-29088
Byzantine-Robust Federated Learning Framework via a Server-Client Defense Mechanisms
Federated learning (FL), a distributed machine learning (ML) framework, is susceptible to Byzantine attacks since the attacker can manipulate clients’ local data or models to compromise the performance of the global model. There has been a wealth of defenses developed to mitigate the attacks by limiting the impact of malicious models. Nevertheless, the attacker can easily circumvent these approaches that rely solely on a single server-side defense, stemming from the high dimensionality of models and the variety of Byzantine attacks. Therefore, we propose Basalt, a Byzantine-robust FL framework with a server-client joint defense mechanism that enables multiple clients to train a global ML model under Byzantine attacks. On the client side, we design an efficient self-defense approach with model-level penalty loss that restricts local-benign divergence and decreases local-malicious correlation to prevent misclassification. On the server side, we present an efficient defense strategy based on the manifold and maximum clique, further strengthening the FL’s resilience against Byzantine attacks. We provide theoretical guarantees for global model convergence in FL with Byzantine attacks. Our extensive experiments demonstrate that Basalt outperforms existing state-of-the-art works. Especially, it achieves nearly 100% accuracy for detecting malicious clients in nonindependent and identically distributed (Non-IID) MNIST datasets under various Byzantine attacks.
- 5 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. 21 (8) , 6433-6442
A Novel Quantitative Risk Assessment Model for Industrial Control Systems Integrating the Cyber–Physical Domain
The tight cyber–physical coupling in industrial control system (ICS) makes it vulnerable to attacks, where attackers can exploit system vulnerabilities to cross domain boundaries, causing significant losses. A novel quantitative risk assessment framework for ICS is proposed, integrating both the cyber domain and the physical domain to address potential risk assessment issues. First, an entropy-weighted technique for order preference by similarity to ideal solution method introducing triangular fuzzy numbers is proposed to solve the vulnerability index for multiattribute decision making to obtain the a prior probability. Second, the risk propagation mechanism of complex network topology and device interaction under ICS was studied, and the posterior probability model based on susceptible-exposed-infectious-recovered was designed to dynamically capture the propagation characteristics of risks in the network. Finally, the safety loss level is defined based on standard criteria to achieve quantitative risk assessment of ICS. The proposed risk assessment model was experimentally validated on the SWaT testbed, with results confirming its feasibility and effectiveness.
- 5 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. 21 (8) , 6433-6442
A Novel Quantitative Risk Assessment Model for Industrial Control Systems Integrating the Cyber–Physical Domain
The tight cyber–physical coupling in industrial control system (ICS) makes it vulnerable to attacks, where attackers can exploit system vulnerabilities to cross domain boundaries, causing significant losses. A novel quantitative risk assessment framework for ICS is proposed, integrating both the cyber domain and the physical domain to address potential risk assessment issues. First, an entropy-weighted technique for order preference by similarity to ideal solution method introducing triangular fuzzy numbers is proposed to solve the vulnerability index for multiattribute decision making to obtain the a prior probability. Second, the risk propagation mechanism of complex network topology and device interaction under ICS was studied, and the posterior probability model based on susceptible-exposed-infectious-recovered was designed to dynamically capture the propagation characteristics of risks in the network. Finally, the safety loss level is defined based on standard criteria to achieve quantitative risk assessment of ICS. The proposed risk assessment model was experimentally validated on the SWaT testbed, with results confirming its feasibility and effectiveness.