Collusion attacks, where the server and malicious clients collaborate to bypass gradient source confirmation or tamper with aggregation results, cause a fundamental damage to the training process in federated learning (FL). However, existing verifiable FL frameworks typically adopt a split-verification model-clients can independently validate the correctness of aggregation results, while the server is responsible for confirming the legitimacy of gradient sources. Thus, the collusion attack has emerged as a critical and intractable vulnerability, as it completely collapses this split-verification model, thereby invalidating such verification mechanisms. To tackle this fundamental issue, we propose an innovative dual-verifiable FL framework. Specifically, by leveraging vector commitments, our scheme first integrates both gradient source confirmation and aggregation result verification into a unified framework. Based on this unified design, our scheme implements two targeted strategies to defend against collusion attacks. To prevent collusion-enabled gradient source spoofing, our scheme introduces a semi-trusted verification cluster in place of unreliable server-side validation and embeds an anonymized identity-check strategy to collaboratively confirm gradient source legitimacy. To counter collusion-driven manipulation of gradient aggregation results, our scheme customizes auxiliary verification proofs with a computational one-wayness for client-uploaded gradients. This renders it infeasible for adversaries to tamper with the aggregation result through reverse engineering. Under experiments and security analyses, our scheme achieves reliable dual-verification and robust resistance to collusion attacks. Moreover, it reduces computation and communication overhead by at least 40.83% and 50.47%, respectively, compared to state-of-the-art verifiable FL schemes.

In Incorrect-Plaintext Attacks (IPA), a malicious Data Owner (DO) compromises system security by encrypting incorrect plaintext, while a malicious Data User (DU) undermines system fairness by falsely claiming decryption failure even when the plaintext is valid. Both security and fairness vulnerabilities can be further amplified by adversaries with quantum computing. However, currently, no existing data sharing scheme can detect such quantum IPA, making it unexplored. This paper proposes LICA, a lattice-based IBE (Identity-Based Encryption) scheme with commitment assistance to detect quantum IPA in data sharing. LICA can identify if a ciphertext is encrypted from the specified plaintext and public key. It leverages the hardness of lattice problems to achieve quantum resistance. Commitment binding the encryption random parameters prevents DU from forging valid verification proofs. Then we present a new LICA-based Data Sharing Protocol (LICADS). Upon suspected quantum IPA, a trusted arbitration authority initiates a probabilistic challenge with the DO and collects evidence from the DO and DU. Then the arbitration authority identifies the malicious party using LICA and signatures to apply penalties. LICA is not only quantum-resistant but also formally proven to achieve commitment security and zero-knowledge proof security, which enables privacy protection during the verification process. Meanwhile, LICADS is the first data sharing protocol capable of detecting quantum IPA. We formally prove that LICA satisfies commitment correctness, verification correctness, binding, privacy preservation, unforgeability and pq-IND-sID-CPA security, and that LICADS achieves fairness, ensuring that any malicious party cannot evade punishment while honest parties are never penalized. We evaluate related schemes in terms of computational overhead, communication overhead, smart contract latency, and consensus efficiency. Experimental results show that LICADS incurs computational overhead comparable to existing schemes and roughly an order-of-magnitude higher communication overhead (about 45 KB under the Module-LWE instantiation versus 4–6 KB). We regard this as a worthwhile trade-off, since LICADS is the only scheme that detects quantum IPA.

Nowadays, crowdsourcing has evolved into a cost-efficient and scalable task execution paradigm that benefits both task requesters and workers. Task matching is a crucial crowdsourcing procedure for deciding the task execution quality, but security and privacy concerns arise as the crowdsourcing platform cannot be fully trusted. Existing privacy-aware task-matching schemes are limited to intra-platform central matching in the semi-honest model and coarse-grained keyword/location-based matching over one single attribute. Solutions supporting secure cross-platform and fine-grained task matching in the malicious model are urgently needed. In this paper, we first formally defined BFCrowd, a federated crowdsourcing system built on a consortium blockchain. BFCrowd aggregates multi-platform resources and enables decentralized and reliable cross-platform task matching using smart contracts, in the presence of malicious workers and platforms. Notably, we design a fully secure ciphertext-policy attribute-based encryption scheme with concealed access policies and user-side lightweight decryption, which thoroughly caters to the dual-side privacy demand and resource-limited workers and serves for fine-grained expressive task matching over multiple attributes. Moreover, it supports comparison over numerical attributes. Formal security analysis proves the desirable privacy guarantees in the standard model and collusion resistance. Extensive experiments implemented atop Hyperledger Fabric demonstrate both on-chain and off-chain performance.

Federated Learning (FL) has recently gained prominence in the context of Vehicular Ad-hoc Networks (VANETs) as a promising approach to enhancing autonomous driving capabilities. However, vehicles' high mobility, real-time communication, and dynamic network topology lead to frequent disconnections during operation, which may slow down the convergence of FL or even lead to training failure. In this study, we propose a scalable aggregation scheme (SAPE) designed to improve computation efficiency and address vehicle dropouts. SAPE employs a lossless encoding algorithm with parallel technology for efficient aggregation of large vectors. Then, we leverage TJL (ACSAC '22) to reconstruct gradients for dropout vehicles, using online vehicles to establish a k-regular graph. In a network with N vehicles, SAPE achieves a secure aggregation overhead of O(log(2 )(N)), as opposed to O(N-2), tolerating a vehicle dropout rate of up to 33%. Furthermore, we conduct a theoretical security analysis of SAPE to prove its security under honest-but-curious (HBC) and malicious attack models. Extensive experiments show that SAPE outperforms existing baseline aggregation schemes by up to 1.4 & times; speedups in aggregation time.

Vehicle-to-infrastructure (V2I) communication connects vehicles and transportation infrastructures (e.g., road-side units, 5G base stations), facilitating bidirectional information exchanges. As a matter of fact, V2I communication is constrained by the limited wireless coverage, which forces vehicles to undergo cross-zone handovers frequently. Existing protocols for cross-zone handover either incur high overhead or neglect the risks of attacks during service disconnections. To tackle these concerns, we propose a seamless cross-zone authentication scheme tailored for V2I communication, called SCZ-HA. Drawing inspiration from the zone encryption scheme (EuroS&P 2020), we integrate long-term and short-term credentials to optimize efficiency and enhance security. When entering a new zone, vehicles use their valid long-term credentials to request short-term credentials from the local infrastructure. These short-term credentials are then encapsulated and distributed to neighboring zones using Shamir's secret sharing, thereby reducing communication overhead. During cross-zone handovers, vehicles only present encapsulations for verification by the target infrastructure, thus eliminating the need to validate full credential contents. Unlike previous schemes, SCZ-HA avoids redundant verifications, enabling millisecond-level handovers and supporting traceability of malicious vehicles. Experimental results show that our scheme achieves a substantial reduction in authentication latency, with an average improvement of 70.54% compared to state-of-the-art schemes.

A decentralized storage network (DSN) collects idle storage resources from Internet nodes for low-cost rental to users, and its scale has grown exponentially. In a DSN, most users rely on a centralized third-party service provider (SP) to process user-side data and interact with decentralized storage nodes (SNs), making users suffer from a single point of failure. Additionally, since both SP and SNs may offer malicious services, enabling fault tolerance, data confidentiality, and availability guarantee with public verifiability is crucial in the presence of such threats. In this paper, we propose TolerStore, a completely decentralized service framework for DSNs with decentralized SPs and SNs, which can tolerate Byzantine SPs and malicious SNs. To the best of our knowledge, TolerStore is the first to develop a blockchain with multiple SPs for privacy-aware data processing in DSNs, which is formally proven to ensure Byzantine fault tolerance, data confidentiality, public verification, and data availability. Furthermore, we propose an optimized Byzantine Fault-Tolerant consensus with an adaptive leader rotation, incorporating homomorphic fingerprints to verify privacy-aware data processing with enhanced performance. We implement a TolerStore prototype over Hyperledger Fabric, and extensive experiments show that it tolerates & LeftFloor;(N-1/)3 & RightFloor; Byzantine SPs and 50% malicious SNs with up to 99.99% data availability.

In vehicular networks, pseudonyms are a fundamental mechanism for achieving anonymous communication. However, existing authentication and key agreement (AKA) schemes typically require each pseudonym to be cryptographically bound to private information, such as a certificate or a secret key, to ensure authenticity. This approach leads to significant challenges, including complex certificate management and the inherent risks of secret key escrow. Furthermore, to maintain unlinkability, vehicles must store a large pool of pseudonyms and their associated private information-leading to prohibitive storage costs. The high computation and communication costs of such schemes are also ill-suited for the delay-sensitive nature of vehicular environments. To address these limitations, we propose a pseudonym-only approach for anonymous AKA in vehicular networks. The primary contribution of our scheme is its novel "pseudonym-only" approach to authentication, without the need to combine it with other private information. PidTree provides an efficient pseudonym generation method. Our scheme involves lightweight computation operations such as hash functions and Lagrange interpolation. The security analysis shows that our scheme satisfies the essential security and privacy requirements of vehicular networks. Our scheme reduces the storage cost for the trusted party from O(MN) to O(M) and the storage cost for a vehicle by at least 86.50%. The performance analysis also shows that our scheme outperforms the representative schemes in terms of computation cost and simulation results.

Fuzzing is a widely used technique for detecting vulnerabilities in smart contracts, which generates transaction sequences to explore the execution paths of smart contracts. However, existing fuzzers are falling short in detecting sophisticated vulnerabilities that require specific attack transaction sequences with proper inputs to trigger, as they (i) prioritize code coverage over vulnerability discovery, wasting considerable effort on non-vulnerable code regions, and (ii) lack semantic understanding of stateful contracts, generating numerous invalid transaction sequences that cannot pass runtime execution. In this paper, we propose SmartFuzz, a novel collaborative reflective fuzzer for smart contract vulnerability detection. It employs large language model-driven agents as the fuzzing engine and continuously improves itself by learning and reflecting through interactions with the environment. Specifically, we first propose a new Continuous Reflection Process (CRP) for fuzzing smart contracts, which reforms the transaction sequence generation as a self-evolving process through continuous reflection on feedback from the runtime environment. Then, we present the Reactive Collaborative Chain (RCC) to orchestrate the fuzzing process into multiple sub-tasks based on the dependencies of transaction sequences. Furthermore, we design a multi-agent collaborative team, where each expert agent is guided by the RCC to jointly generate and refine transaction sequences from both global and local perspectives. We conduct extensive experiments to evaluate SmartFuzz's performance on real-world contracts and DApp projects. The results demonstrate that SmartFuzz outperforms existing state-of-the-art tools: (i) it detects 5.8%-74.7% more vulnerabilities within 30 minutes, and (ii) it reduces false negatives by up to 80%.

Deep learning-based detectors have been widely proposed to predict vulnerabilities in smart contracts, yet their unreliable predictions pose severe security risks to financial transactions, making it critical to verify the reliability of vulnerability predictions. However, existing methods only produce prediction results, failing to provide an evidence chain to check whether these predicted vulnerabilities genuinely exist and deliver further guidance for fixing the vulnerabilities. Thus, making these vulnerability predictions verifiable remains an unexplored problem. In this paper, we propose SmartGuard, a novel verifiable vulnerability prediction framework for deep learning-based detectors and specifically designed for smart contracts. It integrates a deep learning-based detector with a symbolic prediction validator, where the latter acts as the backend formal engine to verify vulnerability predictions. Specifically, we present a graph-sequence multi-task learning model to detect vulnerabilities while generating transaction sequences that serve as evidence chains, explicitly revealing the triggering logic behind vulnerabilities. To bridge the gap between deep learning-based detectors and symbolic validators, we symbolically execute the generated transaction sequences against the verification conditions of vulnerability predictions. Furthermore, we propose a new metric, Vulnerability Prediction Suspiciousness (VPS), to evaluate the reliability of the predicted results. We implement SmartGuard on three representative types of vulnerabilities (Reentrancy, Ether-leaking, and Suicidal) to evaluate its performance in real-world scenarios. Our experimental results show that SmartGuard can effectively verify doubtful vulnerability predictions in real-world scenarios. It also outperforms state-of-the-art baselines by consistently reducing false reports by at least 15% across various Solidity versions. Case studies on complex contracts and DApps further demonstrate SmartGuard’s effectiveness in practice.

Fog-assisted cloud storage enables efficient collection and management of Internet of Things data, while large-scale data raise severe requirements for storage space. Deduplication schemes over similar data have been investigated to relieve the storage pressure. However, existing schemes are designed based on an idealized assumption that users can accept a certain degree of data loss or the stored data modification. When the uploaded data reaches the preset similarity threshold, only one copy will be stored, which causes data loss. Meanwhile, different dynamic operations on only one copy will cause the stored data modification. In this article, we propose U'Dedup to address the above challenge, which is the first block-level deduplication scheme over similar data. The key component of U'Dedup is a self-built tree data structure that supports different update requirements to avoid data modification without duplicating the stored data. U'Dedup ensures the completeness of all unique data to avoid data loss and constructs a dual deduplication architecture to relieve the computing pressure of cloud. Finally, the security analysis proved that U'Dedup is secure in the random oracle model. Experimental results show that U'Dedup achieves 57.3%-90.6% upload computation cost saving, and at least 9.5x retrieval computation cost reduction.

Inserted data redaction is an important and hot security topic for permissioned blockchain. However, existing redactable schemes have two problems: (i) Uncaptured Change: Transaction-level redaction schemes modify both user-inserted data and system-generated transaction data, making it difficult to capture the changes in system-generated transactions data; (ii) Unexpected Spread: When the redaction process deletes harmful data, the consensus process may take a long time, resulting in unexpected spread of harmful data. Instead of eliminating the influence of the harmful data, it promotes its spread. To tackle these problems, we propose an Extended Merkle Tree (EMT) structure for inserted data redaction in permissioned blockchain. EMT separates the system-generated transaction data from user-inserted data, ensuring that only the inserted data can be modified during redaction. Furthermore, we design an EMT-based blockchain ledger and develop a smart contract to implement the inserted data redaction. We prove that our redaction scheme preserves the core security features of blockchain and implements redaction operations in Hyperledger Fabric. The experimental results show that our EMT-based blockchain scheme is capable of redacting the harmful inserted data while incurring less than 5% overhead compared to current permissioned blockchains.

Federated edge learning (FEL) has become a key technology due to its privacy protection for clients. Since during the FEL process, there always exists the parameter passing between edge clients and server under an open communication environment, the learning performance depends heavily on the wireless channel conditions. In this paper, we investigate the performance optimization of FEL system in a practical Internet of Things (IoT) scenario where the channel state information (CSI) is imperfect. A non-convex joint optimization problem for client selection and resource allocation is first built to balance the total energy consumption and learning accuracy of the FEL system. Then for solving the built optimization problem with mixed integer properties, two subproblems are derived by relaxing and dividing. For the resource allocation subproblem, a resource allocation algorithm based off-policy optimization (RAOPO) is proposed to obtain the resource allocation scheme. Based on the resource allocation, an energy-efficient and low-latency client selection algorithm (ELCS) is further designed for improving the performance. The extensive simulations verify that, when considering the imperfect CSI, our proposed ELCS can ensure the learning accuracy and system stability with a low energy cost, which supports the fast development of FEL.

Blockchain with the Internet of Things (BIoT) denotes that the blockchain system is used for managing node identities in centerless IoT. Large-scale BIoT (L-BIoT) means many blockchains are used for managing node identities from different IoTs. The verifiable query scheme for BIoT was widely discussed in real-time IoT system. However, the query scheme will require cross-chain operations when BIoT expands into L-BIoT. Cross-chain verifiable query schemes for BIoT exist two issues. 1) Multiround Consensus: Cross-chain query requires the relay chain and related chains to participate in consensus, and each interaction requires initiating one consensus round. 2) High-Overhead Verification: Each data item returned by various blockchains requires individual integrity verification. We propose CVCQ, an off-chain committee (OC)-based verifiable cross-chain query scheme for L-BIoT. CVCQ achieves only one-round consensus by shifting the on-chain data query and consensus to an OC. Meanwhile, the aggregation accumulator and proofs are utilized to verify data integrity, ensuring the low-overhead verification. We proved that CVCQ reduces complexity of consensus from O(m(lambda+1) + m N-lambda) to O(m(lambda)), and decreases the computational complexity of verification from O(m log n) to O(1), when cross-chain data query involving m regional IoTs with N nodes and n data. The experiment results show that CVCQ outperforms the existing schemes by achieving higher on-chain throughput and lower latency.

Space-aerial-ground collaborative UAV network (SAGUN) is a promising emerging network, and our research focuses on enabling secure and efficient data sharing in SAGUN by utilizing blockchain technology. Given that UAVs and other agencies are concerned about the economic rewards and privacy risks of data sharing, reasonable incentives are essential. However, existing works lack adaptivity to dynamic network environments and accountability for data services. In this paper, we propose an adaptive and accountable incentive scheme for data sharing for SAGUN. First, we design a blockchain-enabled novel architecture for SAGUN, where blockchain serves as a secure base for trusted accountability of SAGUN data sharing. Second, according to the different capabilities of the network to provide information for the formulation of sharing incentive strategies, simultaneously motivating both data buyers and sellers to participate in the sharing, we implement two types of ‘tailor-made’ auction protocols based on smart contracts. 1) The network supports effective synchronization of relevant information for strategy formulation: we implement a sealed-bid reverse auction protocol, which maximizes the buyers’ interests while protecting the sellers’ interests through optimal bidding; 2) The network fails to support effective synchronization of relevant information: we implement an auction protocol based on distributed reinforcement learning, which largely mitigates the information dependency. Finally, compared to existing representative works, the experimental results confirm that the proposed scheme securely and effectively improves the quality of data sharing with low computational and communication overheads.

The demand for cross-chain transactions is increasing with the emergence of different types of cryptocurrencies. Most of solutions require cross-chain transaction information to be collected publicly, which undermines the privacy of cross-chain payers. To address this challenge, we propose a privacy-preserving scheme for cross-chain cryptocurrency transactions, which prevents any party from accessing the full transaction details of the payers. Based on the commitment scheme, we first construct a privacy-preserving cryptocurrency lock and release protocol for issuing or redeeming mapped cryptocurrencies in smart contracts. We then propose a mapped cryptocurrency confidential transaction scheme (MCCT) based on zk-SNARK and design an atomic exchange protocol based on it. We analyze the security of our scheme and implement it with smart contracts on an Ethereum test network. The experimental results demonstrate that for 100 exchanges, the additional time cost of our proposed solution is approximately 39% compared to the existing solution, which is considered acceptable.

Electronic auctions (e-auctions) remove the physical limitations of traditional auctions and bring this mechanism to the general public. However, most e-auction schemes involve a trusted auctioneer, which is not always credible in practice. Some studies have applied cryptography tools to solve this problem by distributing trust, but they ignore the existence of collusion. In this paper, a blockchain-based Privacy-Preserving and Collusion-Resistant scheme (PPCR) for double auctions is proposed by employing both cryptography and blockchain technology, which is the first decentralized and collusion-resistant double auction scheme that guarantees bidder anonymity and bid privacy. A two-server-based auction framework is designed to support off-chain allocation with privacy preservation and on-chain dispute resolution for collusion resistance. A Dispute Resolution agreement (DR) is provided to the auctioneer to prove that they have conducted the auction correctly and the result is fair and correct. In addition, a Concise Dispute Resolution protocol (CDR) is designed to handle situations where the number of accused winners is small, significantly reducing the computation cost of dispute resolution. Extensive experimental results confirm that PPCR can indeed achieve efficient collusion resistance and verifiability of auction results with low on-chain and off-chain computational overhead.

Cybercriminals pose a significant threat to blockchain trading security, causing $40.9 billion in losses in 2024. However, the lack of an effective real-world address dataset hinders the advancement of cybercrime detection research. The anti-cybercrime efforts of researchers from broader fields, such as statistics and artificial intelligence, are blocked by data scarcity. In this paper, we present Real-CATS, a Real-world dataset of Cryptocurrency Addresses with Transaction profileS, serving as a practical training ground for developing and assessing detection methods. Real-CATS comprises 103,203 criminal addresses from real-world reports and 106,196 benign addresses from exchange customers. It satifies the C3R characteristics (Comprehensiveness, Classifiability, Customizability, and Real-world Transferability), which are fundemental for practical detection of cryptocurrency cybercrime. The dataset provides three main functions: 1) effective evaluation of detection methods, 2) support for feature extensions, and 3) a new evaluation scenario for real-world deployment. Real-CATS also offers opportunities to expand cybercrime measurement studies. It is particularly beneficial for researchers without cryptocurrency-related knowledge to engage in this emerging research field. We hope that studies on cryptocurrency cybercrime detection will be promoted by an increasing number of cross-disciplinary researchers drawn to this versatile data platform. All datasets are available at https://github.com/sjdseu/Real-CATS

Multi-chain data sharing refers to cross-chain data exchange among multiple blockchains. However, existing multi-chain data sharing schemes rely on direct blockchain-to-blockchain connections to establish links among multiple chains. This leads to poor scalability and low efficiency as the number of connected blockchains increases. To address these problems, we propose GAM (Group Authorization-based Multi-chain Data Sharing) for scalable and efficient multi-chain data sharing. GAM enhances scalability by organizing users from different chains into authorized virtual groups, enabling trusted data sharing within it. To further improve efficiency, the data sharing authorization process is executed on-chain, while data transfer is based on off-chain storage. We provide a formal analysis of GAM in multi-chain scenarios and implement a proof-of-concept prototype using Hyperledger Fabric. Experimental results demonstrate that GAM is effective in reducing the execution time of multi-chain data sharing while maintaining high transaction throughput and minimal end-to-end delay.

As a critical component in federated learning (FL), secure aggregation enables the server to learn the aggregated model without observing clients' local training gradients. However, limited by computation and communication capabilities, existing aggregation schemes are not suitable to be directly employed in the Vehicular Ad Hoc Networks (VANETs) scenario. In this paper, we present a secure aggregation framework constructed with k-regular graph over VANETs scenario. We first optimize the secure aggregation scheme proposed by Bell et al. (CCS 2020). Specifically, using this new building block and an identity authentication mechanism in the vehicle-to-vehicle (V2V) communication mode, we design an optimized aggregation scheme that, when executed among n vehicles, can further reduce $2n$ communication times between vehicles and the central server while guaranteeing logarithmic overhead. Besides, by applying a zero-knowledge proof to the authentication process, our proposal supports vehicles anonymously constructing the k-regular graph and completing parameter computation process, which enhances privacy preservation in semi-honest settings. Under the experiment and security analysis, our proposal is demonstrated to be able to effectively achieve privacy preservation while achieving less computation and communication overheads compared to state-of-the-art aggregation schemes.