Wenyuan Yang

Shuo Shao, Wenyuan Yang, Hanlin Gu et al.

Federated learning (FL) is a distributed machine learning paradigm allowing multiple clients to collaboratively train a global model without sharing their local data. However, FL entails exposing the model to various participants. This poses a risk of unauthorized model distribution or resale by the malicious client, compromising the intellectual property rights of the FL group. To deter such misbehavior, it is essential to establish a mechanism for verifying the ownership of the model and as well tracing its origin to the leaker among the FL participants. In this paper, we present FedTracker, the first FL model protection framework that provides both ownership verification and traceability. FedTracker adopts a bi-level protection scheme consisting of global watermark mechanism and local fingerprint mechanism. The former authenticates the ownership of the global model, while the latter identifies which client the model is derived from. FedTracker leverages Continual Learning (CL) principles to embed the watermark in a way that preserves the utility of the FL model on both primitive task and watermark task. FedTracker also devises a novel metric to better discriminate different fingerprints. Experimental results show FedTracker is effective in ownership verification, traceability, and maintains good fidelity and robustness against various watermark removal attacks.

Wenyuan Yang, Shuo Shao, Yue Yang et al.

Federated learning (FL) allows multiple participants to collaboratively build deep learning (DL) models without directly sharing data. Consequently, the issue of copyright protection in FL becomes important since unreliable participants may gain access to the jointly trained model. Application of homomorphic encryption (HE) in secure FL framework prevents the central server from accessing plaintext models. Thus, it is no longer feasible to embed the watermark at the central server using existing watermarking schemes. In this paper, we propose a novel client-side FL watermarking scheme to tackle the copyright protection issue in secure FL with HE. To our best knowledge, it is the first scheme to embed the watermark to models under the Secure FL environment. We design a black-box watermarking scheme based on client-side backdooring to embed a pre-designed trigger set into an FL model by a gradient-enhanced embedding method. Additionally, we propose a trigger set construction mechanism to ensure the watermark cannot be forged. Experimental results demonstrate that our proposed scheme delivers outstanding protection performance and robustness against various watermark removal attacks and ambiguity attack.

Zhen Sun, Zongmin Zhang, Leyi Sheng et al.

Image steganography is widely used to protect user privacy and enable covert communication. However, it can also be abused by the adversary as a covert channel to bypass content moderation, disseminate harmful semantics, and even hide malicious instructions in images to elicit dangerous outputs from large models, posing a practical security risk that continues to evolve. To address the lack of a unified and systematic evaluation framework, we propose SADBench, a systematic benchmark that assesses the adversary's ability to inject harmful secrets via steganography and the defender's ability to detect such threats through steganalysis. Crucially, SADBench comprises $4$ core tasks, namely steganography attack capability evaluation, steganalysis defense capability evaluation, efficiency evaluation, and transferability evaluation. It evaluates both image-payload and text-payload steganography across diverse cover distributions, utilizing harmful visual semantics and toxic instructions to simulate malicious attacks. Across a broad set of attacks and detectors, SADBench reveals that (i) INN and autoencoder-based methods demonstrate superior stability compared to other architectures, (ii) in-domain detection is near-perfect and cheaper than generation, (iii) a critical asymmetry exists in transferability where attacks robustly generalize to new distributions while detectors fail to adapt, and (iv) real-world threats persist on social media, where payloads either survive minimal compression or effectively adapt to aggressive compression via simulated training. Overall, SADBench establishes a systematic, reproducible, and extensible framework to quantify risks, paving the way for measurable and security-driven advancements in steganography defense.

Yuguo Yin, Yuxin Xie, Wenyuan Yang et al.

Multilingual audio-text retrieval (ML-ATR) is a challenging task that aims to retrieve audio clips or multilingual texts from databases. However, existing ML-ATR schemes suffer from inconsistencies for instance similarity matching across languages. We theoretically analyze the inconsistency in terms of both multilingual modal alignment direction error and weight error, and propose the theoretical weight error upper bound for quantifying the inconsistency. Based on the analysis of the weight error upper bound, we find that the inconsistency problem stems from the data distribution error caused by random sampling of languages. We propose a consistent ML-ATR scheme using 1-to-k contrastive learning and audio-English co-anchor contrastive learning, aiming to mitigate the negative impact of data distribution error on recall and consistency in ML-ATR. Experimental results on the translated AudioCaps and Clotho datasets show that our scheme achieves state-of-the-art performance on recall and consistency metrics for eight mainstream languages, including English. Our code will be available at https://github.com/ATRI-ACL/ATRI-ACL.

Junpei Liao, Zhikai Chen, Liang Yi et al.

Deep learning based video frame interpolation (VIF) method, aiming to synthesis the intermediate frames to enhance video quality, have been highly developed in the past few years. This paper investigates the adversarial robustness of VIF models. We apply adversarial attacks to VIF models and find that the VIF models are very vulnerable to adversarial examples. To improve attack efficiency, we suggest to make full use of the property of video frame interpolation task. The intuition is that the gap between adjacent frames would be small, leading to the corresponding adversarial perturbations being similar as well. Then we propose a novel attack method named Inter-frame Accelerate Attack (IAA) that initializes the perturbation as the perturbation for the previous adjacent frame and reduces the number of attack iterations. It is shown that our method can improve attack efficiency greatly while achieving comparable attack performance with traditional methods. Besides, we also extend our method to video recognition models which are higher level vision tasks and achieves great attack efficiency.

Wenyuan Yang, Gongxi Zhu, Yuguo Yin et al.

Federated learning allows multiple parties to collaborate in learning a global model without revealing private data. The high cost of training and the significant value of the global model necessitates the need for ownership verification of federated learning. However, the existing ownership verification schemes in federated learning suffer from several limitations, such as inadequate support for a large number of clients and vulnerability to ambiguity attacks. To address these limitations, we propose a cryptographic signature-based federated learning model ownership verification scheme named FedSOV. FedSOV allows numerous clients to embed their ownership credentials and verify ownership using unforgeable digital signatures. The scheme provides theoretical resistance to ambiguity attacks with the unforgeability of the signature. Experimental results on computer vision and natural language processing tasks demonstrate that FedSOV is an effective federated model ownership verification scheme enhanced with provable cryptographic security.

Wenyuan Yang, Yuguo Yin, Gongxi Zhu et al.

Federated learning (FL) allows multiple parties to cooperatively learn a federated model without sharing private data with each other. The need of protecting such federated models from being plagiarized or misused, therefore, motivates us to propose a provable secure model ownership verification scheme using zero-knowledge proof, named FedZKP. It is shown that the FedZKP scheme without disclosing credentials is guaranteed to defeat a variety of existing and potential attacks. Both theoretical analysis and empirical studies demonstrate the security of FedZKP in the sense that the probability for attackers to breach the proposed FedZKP is negligible. Moreover, extensive experimental results confirm the fidelity and robustness of our scheme.

Guoxiong Su, Wenyuan Yang, Zhengding Luo et al.

The need for data trading promotes the emergence of data market. However, in conventional data markets, both data buyers and data sellers have to use a centralized trading platform which might be dishonest. A dishonest centralized trading platform may steal and resell the data seller's data, or may refuse to send data after receiving payment from the data buyer. It seriously affects the fair data transaction and harm the interests of both parties to the transaction. To address this issue, we propose a novel blockchain-based data trading framework with Trusted Execution Environment (TEE) to provide a trusted decentralized platform for fair data trading. In our design, a blockchain network is proposed to realize the payments from data buyers to data sellers, and a trusted exchange is built by using a TEE for the first time to achieve fair data transmission. With these help, data buyers and data sellers can conduct transactions directly. We implement our proposed framework on Ethereum and Intel SGX, security analysis and experimental results have demonstrated that the framework proposed can effectively guarantee the fair completion of data tradings.