LHLijuan Huo

Lijuan Huo

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Lijuan Huo has a total of 29 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Heterogeneous Privacy-Preserving Federated Learning for Edge Intelligence

    by Zhuangzhuang Zhang, Libing Wu, Helei Cui, Zhibo Wang, Lijuan Huo, Jing Wang, Shengshan Hu

    Federated learning (FL) as a distributed machine learning paradigm can be applied to edge intelligence scenarios for collaborative machine learning model building. Unfortunately, existing privacy-preserving FL applied to this scenario still faces three challenges: data heterogeneity, model heterogeneity, and privacy heterogeneity. Despite numerous privacy-preserving FL techniques proposed, they still cannot effectively address these three challenges. To solve this problem, we propose HeteroFed, a heterogeneous privacy-preserving FL framework for edge intelligence. Our HeteroFed contains heterogeneous model construction, dynamic gradient clipping, adaptive noise addition, and deviation-aware model aggregation. Specifically, we first use the heterogeneous model construction mechanism to enable personalized model training for different smart devices. Then, we propose a dynamic gradient clipping mechanism to perform dynamically adjusted gradient clipping on models uploaded by smart devices to limit the magnitude of gradients. Finally, we propose an adaptive noise addition mechanism to customize differential privacy protection for smart device models based on their convergence status. Furthermore, to mitigate the influence of noise perturbations on model performance, we propose a deviation-aware model aggregation mechanism for accurate model aggregation. Theoretical analysis demonstrates that HeteroFed achieves heterogeneous differential privacy. Extensive experiments show that HeteroFed outperforms similar methods, improving global model accuracy by 18%, 15%, 13%, and 18% on the MNIST, Fashion-MNIST, CIFAR-10, and THUCNews datasets, respectively.


  • Transferable Adversarial Attacks on 3-D Point Cloud Semantic Segmentation via Diffusion Models in Autonomous Driving

    by Yizhou Wang, Libing Wu, Zhuangzhuang Zhang, Lijuan Huo, Jiaqi Feng, Jing Wang, Jiong Jin

    With the rapid advancement of LiDAR-based 3D semantic segmentation, it has provided effective support for comprehensive perception in autonomous driving. However, the vulnerabilities of existing segmentation models have not been fully explored, posing potential risks to the security of autonomous systems. Existing adversarial attacks on point cloud segmentation primarily focus on applying global perturbations to point coordinates, inevitably introducing noticeable outliers and resulting in suboptimal attack performance. Inspired by the remarkable performance of diffusion models in LiDAR scene completion tasks, we propose a novel adversarial attack framework using diffusion models targeting 3D semantic segmentation. First, we model forward and reverse diffusion processes for input point cloud scenes, then introduce targeted perturbations during the reverse diffusion sampling stage. Specifically, we initialize Gaussian noise and add it to partial input point clouds, followed by stepwise denoising for point cloud reconstruction. During each denoising step, we identify the most discrimination-sensitive critical points and apply adversarial gradients exclusively to these selected points for guiding updates. Furthermore, we design a triple-loss function to supervise adversarial point cloud generation, achieving a balance between attack performance and perturbation magnitude. Finally, we conduct comprehensive experiments on two autonomous driving datasets, SemanticKITTI and nuScenes. The experimental results indicate that our method delivers significant attack performance against six mainstream semantic segmentation models, with average relative mIoU drops reaching 61.14% and 51.55% respectively. Moreover, the proposed method successfully transfers to point cloud classification and part segmentation tasks, demonstrating superior attack performance and transferability.


  • Efficient Dynamic Auditing Scheme for Cloud Data Sharing of Intelligent Connected Fleet

    by Jiaqi Feng, Libing Wu, Lijuan Huo, Zhuangzhuang Zhang, Xing Fan, Jiong Jin

    Intelligent connected vehicle fleet driven primarily by cloud data sharing is an efficient and energy-saving organizational model and demonstrates vast application potential in many scenarios. However, as the central hub for data sharing, the cloud platform's untrustworthiness poses severe challenges to the integrity and reliability of shared data. Existing auditing schemes fail to meet the low-latency, lightweight onboard computing, and succinct revocation demands of such fleets. Even with the application of batch auditing, the overall audit latency remains inadequately addressed. In this work, we propose an efficient auditing scheme called Drivora for cloud data sharing of intelligent connected fleet. We achieve secure delegated authenticator computation based on secret sharing and it requires only the master vehicle to perform the initial setup. Moreover, Drivora supports the aggregation of challenged blocks from multiple vehicles to produce a single integrity proof and it enables Drivora to handle multiple audit tasks from different vehicles at the same time. Furthermore, we propose a novel vehicle revocation mechanism. Different from traditional secret sharing-based approaches, Drivora requires only a single secret share to revoke multiple vehicles and achieves lower space complexity. We conduct extensive experimental evaluations of Drivora's performance and benchmark it against state-of-the-art approaches. Compared to conventional batch auditing, our scheme achieves a $2\times$-$107\times$ improvement in auditing efficiency and reduces communication overhead by $2\times$-$939\times$. As the number of files increases, our overall auditing cost remains constant.

  • SecChain: A Secure Stateless Sharded Blockchain via a Novel State Commitment

    by Lijuan Huo, Libing Wu, Enshu Wang, Bolong Zheng, Xing Fan, Xinhai Yan, Bingyi Liu

    Sharding protocols in cryptocurrencies are dedicated to improving system throughput and reducing cross-shard transaction latency. However, they face the challenge of state explosion due to their stateful design. An intuitive solution is state sharding. Existing state sharding protocols aim to reduce the incidence of cross-shard transactions and balance transaction load. Nevertheless, it inevitably introduces the complex and sequential processing of cross-shard transactions and neglects the security of state data. To address these challenges, we propose SecChain, a secure stateless sharded blockchain via a novel state commitment in this paper. SecChain fully exploits the historical characteristics of transactions and designs the parallel cross-shard transaction execution based on asynchronous prepaid accounts. We then introduce a novel state commitment scheme to ensure the security of state data and improve the efficiency of transaction validation. In addition, we optimize the on-chain and off-chain storage by integrating an incremental partial state trie on-chain and a full trie with proofs off-chain. We perform a comprehensive evaluation using real-world workloads and results demonstrate that SecChain's throughput is 1.75× that of BrokerChain, and its latency is reduced by 7×. Furthermore, SecChain significantly reduces both proof overhead and storage overhead.


  • Open Access

    FedBAP: Backdoor Defense via Benign Adversarial Perturbation in Federated Learning

    by Xinhai Yan, Libing Wu, Zhuangzhuang Zhang, Bingyi Liu, Lijuan Huo, Jing Wang


  • Ascina: Efficient Proof of Retrievability for Industrial Cloud Storage Systems

    by Jiaqi Feng, Libing Wu, Xing Fan, Lijuan Huo, Enshu Wang, Zhuangzhuang Zhang

    Industrial cloud storage systems enhance data availability and offer intelligent services to enterprises. However, they also bring significant concerns about data integrity since the cloud space provider may not consistently retain the outsourced data. Existing cloud storage verification methods impose a significant computational burden on edge devices, so they are unsuitable for industrial cloud storage systems. Although the light-weight homomorphic authenticator somewhat alleviates the computational load on the fog node, its effectiveness remains limited and it further complicates proof generation and verification. To address these problems, we propose Ascina, a new proof of retrievability framework for industrial cloud storage systems. We employ verifiable secret sharing to delegate tag computation tasks to the computing server without disclosing the signing key. Once the fog node completes its initial configuration, no further computations are required. Additionally, we propose the improved Ascina by utilizing Fast Fourier Transform (FFT) and Inverse Fast Fourier Transform (IFFT) technology to further alleviate the computational burden on the fog node. Furthermore, we propose a batch verification algorithm that simultaneously validates the integrity of multiple files while maintaining the same security assurances as single auditing. We evaluate the performance of Ascina through experiments and compare it with state-of-the-art methods. Experimental results demonstrate that the computational overhead on the fog node in Ascina is 5×-251× lower than that of Edasvic and our proof verification time and proof size are reduced by 5×-16× and 977×-3285×, respectively. As the size and quantity of files grow, Ascina demonstrates greater efficiency in both time and space.


  • NTTproofs: A Maintainable and Aggregatable Vector Commitment With Fast Openings and Updates

    by Lijuan Huo, Libing Wu, Enshu Wang, Jinfei Liu, Chunshuo Li, Zeimei Liu, Zhuangzhuang Zhang

    In vector commitments, the complex process of generating and updating proofs, along with the large-sized proofs, seriously hinders the practicality of stateless cryptocurrencies. In this work, we present NTTproofs, containing two sub-schemes, a vector commitment (VC) and a mulit-vector commitment (MC). Both sub-schemes are maintainable and aggregatable, and they also enjoy fast openings (i.e., generating all the proofs) as well as efficient proof updates. MC in NTTproofs employs the Fast Number Theoretic Transform (NTT) and sharding technique to significantly improve the time of generating all proofs by up to 0.76× and 0.32×, respectively, over Balanceproofs, Matproofs. Moreover, our proposed MC in NTTproofs is efficiently maintainable and requires merely 15.78 milliseconds at n1 = n2 = 212 to update all proofs. Meanwhile, NTTproofs schemes exhibit superior aggregatability, taking 0.003 seconds in VC and 0.05 seconds in MC to aggregate 1024 proofs and reducing the size of an aggregated proof to a constant size of 96 Bytes. Finally, macrobenchmarks indicate that our proposed MC in NTTproofs outperforms the other schemes, but is slightly inferior to that of Balanceproofs.