Selected work

Representative Papers

Semi-Supervised Noise Adaptation: Transferring Knowledge from Noise Domain

May 30, 2026

This work proposes the semi-supervised noise adaptation (SSNA) problem, which aims to enhance model generalization in semi-supervised settings by leveraging a synthetic noise domain—such as Gaussian noise—lacking semantic information as a surrogate source domain, using only a small number of labeled target-domain samples. To address this challenge, the authors introduce the Noise Adaptation Framework (NAF), which, for the first time, incorporates synthetic noise domains into semi-supervised transfer learning and derives a theoretical generalization bound to guide algorithm design. Empirical results demonstrate that NAF effectively exploits knowledge from the noise domain to tighten the generalization bound on the target domain, leading to significant performance improvements over existing semi-supervised learning methods across multiple benchmarks.

0 citationsRead paper

FedRE: A Representation Entanglement Framework for Model-Heterogeneous Federated Learning

Nov 27, 2025

Federated learning (FL) faces concurrent challenges of client model architecture heterogeneity, privacy leakage risks (e.g., representation inversion attacks), and high communication overhead. Method: This paper proposes FedRE, a novel FL framework that (i) generates entangled representations via normalized random weighting to align heterogeneous local feature spaces without architectural alignment; (ii) incorporates label encoding and per-round resampling to mitigate global classifier overconfidence, enhancing generalization and resilience against inversion attacks; and (iii) jointly optimizes local feature aggregation and global supervised training. Contribution/Results: FedRE achieves state-of-the-art efficiency, robustness, and privacy preservation in heterogeneous FL. Experiments on CIFAR-10/100 show ~42% lower communication cost than FedAvg, provable resistance to representation inversion attacks, and accuracy on par with homogeneous FL baselines—marking the first framework to simultaneously achieve all three properties in heterogeneous settings.

0 citationsRead paper

Hashed Watermark as a Filter: Defeating Forging and Overwriting Attacks in Weight-based Neural Network Watermarking

Jul 15, 2025

Deep neural network watermarking (NNW) suffers from poor robustness against forgery and overwrite attacks, particularly for weight-based methods vulnerable to parameter tampering. This paper proposes NeuralMark, the first framework to introduce a hash-based watermark filtering mechanism: an irreversible binary hash watermark acts as a parameter selector, tightly coupling the watermark with model weights; average pooling is further integrated to enhance resilience against fine-tuning and pruning. NeuralMark is architecture-agnostic—compatible with both CNNs and Transformers—and supports diverse tasks, including image classification and text generation. Evaluated across 13 mainstream models, NeuralMark achieves significant improvements in robustness against forgery, overwrite, and compression attacks, while incurring minimal accuracy degradation (<1.2%). A formal security analysis is provided, establishing theoretical guarantees for watermark integrity and unforgeability.

0 citationsRead paper

Noise May Contain Transferable Knowledge: Understanding Semi-supervised Heterogeneous Domain Adaptation from an Empirical Perspective

Feb 19, 2025

This paper addresses the fundamental challenge in semi-supervised heterogeneous domain adaptation (SHDA): knowledge transfer under feature-space and distributional divergence between source and target domains, where only the source domain is fully labeled, the target domain provides only minimal labels, and no instance-level correspondences exist. Through a large-scale empirical study across 330 SHDA tasks, we first reveal that transferable knowledge resides primarily in the *transferability* and *discriminability* of source data—not semantic consistency or structural fidelity—even random noise can serve as effective source data. Building on this insight, we propose the Knowledge Transfer Framework (KTF), a unified approach for knowledge migration. Evaluated on multi-source heterogeneous tasks, KTF achieves performance on par with real source domains and significantly improves target-domain accuracy. The code and datasets are publicly released.

0 citationsRead paper

Seamless Graph Task Scheduling over Dynamic Vehicular Clouds: A Hybrid Methodology for Integrating Pilot and Instantaneous Decisions

Feb 18, 2025

Real-time scheduling of graph-structured computational tasks in dynamic vehicular clouds faces challenges from volatile communication quality, fluctuating computing capacity, uncertain vehicle contact durations, and high data exchange overhead. Method: This paper proposes a low-latency, low-overhead subgraph–vehicle matching framework that jointly optimizes communication reliability, compute volatility, contact uncertainty, and data transfer costs. We introduce a hybrid scheduling paradigm integrating offline risk-aware prior subgraph search (RA-PilotISS) with online real-time isomorphic subgraph search (TE-InstaISS), synergistically balancing predictive capability and response robustness. The method leverages graph isomorphism matching and dynamic risk modeling within a unified vehicular cloud resource management architecture to enable coordinated online/offline decision-making. Results: Extensive experiments across multi-scale scenarios and diverse graph topologies demonstrate significant reductions in task completion time and subgraph search latency; average data exchange cost decreases by 23.7%, outperforming state-of-the-art approaches.

0 citationsRead paper

Recent publications

Latest Papers

Semi-Supervised Noise Adaptation: Transferring Knowledge from Noise Domain

May 30, 2026

This work proposes the semi-supervised noise adaptation (SSNA) problem, which aims to enhance model generalization in semi-supervised settings by leveraging a synthetic noise domain—such as Gaussian noise—lacking semantic information as a surrogate source domain, using only a small number of labeled target-domain samples. To address this challenge, the authors introduce the Noise Adaptation Framework (NAF), which, for the first time, incorporates synthetic noise domains into semi-supervised transfer learning and derives a theoretical generalization bound to guide algorithm design. Empirical results demonstrate that NAF effectively exploits knowledge from the noise domain to tighten the generalization bound on the target domain, leading to significant performance improvements over existing semi-supervised learning methods across multiple benchmarks.

0 citationsRead paper

FedRE: A Representation Entanglement Framework for Model-Heterogeneous Federated Learning

Nov 27, 2025

Federated learning (FL) faces concurrent challenges of client model architecture heterogeneity, privacy leakage risks (e.g., representation inversion attacks), and high communication overhead. Method: This paper proposes FedRE, a novel FL framework that (i) generates entangled representations via normalized random weighting to align heterogeneous local feature spaces without architectural alignment; (ii) incorporates label encoding and per-round resampling to mitigate global classifier overconfidence, enhancing generalization and resilience against inversion attacks; and (iii) jointly optimizes local feature aggregation and global supervised training. Contribution/Results: FedRE achieves state-of-the-art efficiency, robustness, and privacy preservation in heterogeneous FL. Experiments on CIFAR-10/100 show ~42% lower communication cost than FedAvg, provable resistance to representation inversion attacks, and accuracy on par with homogeneous FL baselines—marking the first framework to simultaneously achieve all three properties in heterogeneous settings.

0 citationsRead paper

Hashed Watermark as a Filter: Defeating Forging and Overwriting Attacks in Weight-based Neural Network Watermarking

Jul 15, 2025

Deep neural network watermarking (NNW) suffers from poor robustness against forgery and overwrite attacks, particularly for weight-based methods vulnerable to parameter tampering. This paper proposes NeuralMark, the first framework to introduce a hash-based watermark filtering mechanism: an irreversible binary hash watermark acts as a parameter selector, tightly coupling the watermark with model weights; average pooling is further integrated to enhance resilience against fine-tuning and pruning. NeuralMark is architecture-agnostic—compatible with both CNNs and Transformers—and supports diverse tasks, including image classification and text generation. Evaluated across 13 mainstream models, NeuralMark achieves significant improvements in robustness against forgery, overwrite, and compression attacks, while incurring minimal accuracy degradation (<1.2%). A formal security analysis is provided, establishing theoretical guarantees for watermark integrity and unforgeability.

0 citationsRead paper

Noise May Contain Transferable Knowledge: Understanding Semi-supervised Heterogeneous Domain Adaptation from an Empirical Perspective

Feb 19, 2025

This paper addresses the fundamental challenge in semi-supervised heterogeneous domain adaptation (SHDA): knowledge transfer under feature-space and distributional divergence between source and target domains, where only the source domain is fully labeled, the target domain provides only minimal labels, and no instance-level correspondences exist. Through a large-scale empirical study across 330 SHDA tasks, we first reveal that transferable knowledge resides primarily in the *transferability* and *discriminability* of source data—not semantic consistency or structural fidelity—even random noise can serve as effective source data. Building on this insight, we propose the Knowledge Transfer Framework (KTF), a unified approach for knowledge migration. Evaluated on multi-source heterogeneous tasks, KTF achieves performance on par with real source domains and significantly improves target-domain accuracy. The code and datasets are publicly released.

0 citationsRead paper

Seamless Graph Task Scheduling over Dynamic Vehicular Clouds: A Hybrid Methodology for Integrating Pilot and Instantaneous Decisions

Feb 18, 2025

Real-time scheduling of graph-structured computational tasks in dynamic vehicular clouds faces challenges from volatile communication quality, fluctuating computing capacity, uncertain vehicle contact durations, and high data exchange overhead. Method: This paper proposes a low-latency, low-overhead subgraph–vehicle matching framework that jointly optimizes communication reliability, compute volatility, contact uncertainty, and data transfer costs. We introduce a hybrid scheduling paradigm integrating offline risk-aware prior subgraph search (RA-PilotISS) with online real-time isomorphic subgraph search (TE-InstaISS), synergistically balancing predictive capability and response robustness. The method leverages graph isomorphism matching and dynamic risk modeling within a unified vehicular cloud resource management architecture to enable coordinated online/offline decision-making. Results: Extensive experiments across multi-scale scenarios and diverse graph topologies demonstrate significant reductions in task completion time and subgraph search latency; average data exchange cost decreases by 23.7%, outperforming state-of-the-art approaches.

0 citationsRead paper