Multi-designated verifier signatures (MDVS) are a signer privacy-preserving signature scheme where non-designated parties cannot tell the difference between signatures produced by the true signer and the designated verifiers. However, such privacy-preservation does not allow even the signer to disclose that her signature has been actually created by herself. Thus, it is worth giving to a signer of MDVS claimability, the ability to later make a claim that convinces anyone including non-designated third parties that a signature has been created by the actual signer. In this paper, we propose such a scheme, claimable multi-designated verifier signatures (CMDVS). We define a syntax and security requirements for CMDVS and propose the generic constructions of two types of CMDVS. The first one consists of standard signatures, ring signatures, commitment schemes, and pseudorandom functions (PRF). The second one is the transformation from the first one, using one-time signatures, public key encryption schemes, PRF and non-interactive zero-knowledge proofs.
With the rapid proliferation of wireless devices, ensuring secure communication over open wireless channels has become increasingly critical. Physical layer key generation has emerged as a lightweight and information-theoretically secure cryptographic mechanism, offering a promising complement to traditional cryptographic approaches. However, in Internet of Things (IoT) scenarios characterized by heterogeneous devices and highly dynamic environments, conventional physical-layer key generation methods suffer from low key generation rates, poor stability, and limited adaptability. To address these challenges, this paper proposes a novel AI-based physical layer key generation framework that integrates multi-source signal fusion and intelligent feature selection. Specifically, deep learning models are employed to fuse features from multiple heterogeneous wireless signal sources, enabling the extraction of high-quality randomness and enhancing key entropy and security. In parallel, an attention mechanism is employed to dynamically select the most suitable physical layer features based on real-time environmental conditions, thereby enhancing the system’s adaptability and robustness in complex IoT settings. Finally, we outline potential future research directions and discuss the feasibility of implementing the proposed framework in real-world deployments.
The advent of quantum computing has intensified concerns about the security of widely deployed cryptographic primitives, including RSA and elliptic curve cryptography. In turn, this has heightened demand for next‑generation cryptographic algorithms capable of withstanding quantum attacks, with lattice‑based cryptography emerging as a leading candidate for both encryption and signature schemes. The security of lattice‑based cryptography is grounded in the Shortest Vector Problem (SVP), whose hardness remains unresolved despite numerous proposed algorithms aimed at assessing or attacking it. Among algorithms for tackling SVP, the Sieve family is characterized by a trade‑off between running time and space complexity. In this paper, we introduce the Ideal Triple Sieve, which combines the space efficiency of the Tuple Minkowski Sieve with the rotational advantages of the Ideal Gauss Sieve. The algorithm is designed for broad applicability and, in particular, achieves superior performance on prime cyclotomic ideal lattices, where it outperforms the Ideal Gauss Sieve. Consequently, it yields Minkowski‑reduced lattice vectors whose quality is substantially higher than that of Gauss‑reduced vectors. At the same time, the Ideal Triple Sieve retains the same asymptotic space complexity as the Ideal Gauss Sieve, O(1/n2^0.1887n+o(n)) , and its running time is O(n^22^0.4812n+o(n)) . In addition, we develop a set of techniques that exploit structural properties of ideal lattices to further improve the efficiency of the Ideal Triple Sieve. These techniques enable faster evaluation of norms and inner products and eliminate unnecessary reduction steps, thereby markedly reducing execution time. Finally, we present experimental results for our proposal, demonstrating efficiency compared with existing methods.
Botnet intrusion detection faces several critical challenges, including data silos, significant class imbalance, privacy concerns, and non-independent and identically distributed (Non-IID) client data. To address these challenges, we propose a novel intrusion detection framework based on federated learning. Our framework allows isolated data holders to collaboratively train a shared model while preserving data privacy. To mitigate aggregation bias caused by statistical heterogeneity in the federated setting, we introduce a Jensen-Shannon (JS) divergence-based client similarity-aware aggregation strategy, which dynamically adjusts each client's contribution during global model updates. Furthermore, we integrate the Synthetic Minority Over-sampling Technique (SMOTE) and Focal Loss into the local training process to enhance the model's capability in detecting minority-class attack traffic. Extensive experiments conducted on the publicly available CTU-13 botnet dataset demonstrate that our approach significantly outperforms both conventional centralized models and standard federated learning baselines in terms of accuracy, F1-score, and other key metrics. The results highlight the framework's effectiveness in privacy-preserving modeling of imbalanced and Non-IID intrusion detection data.
Data protection and verification are constantly at risk, and blockchain technology is trying to address this problem effectively. This technology fundamentally changed how we perceive our online safety. Two widely recognized blockchain technology variations namely ethereum and lightweight will be examined in the following paper. The two technologies will be compared with respect to execution time, gas prices, storage efficiency, and scalability. Our research will validate digital data using small to large datasets and propose the best framework for large scale applications. According to our findings, the lightweight blockchain increases the efficiency of gas consumption by 40
The adoption of Large Language Models (LLMs) in businesses raises the possibility of inadvertent intellectual property (IP) and secret data leaks to public artificial intelligence systems. Organizations are using security solutions, including data loss prevention (DLP), access restrictions, monitoring systems, and many more, to lower these risks. However, source code presents a unique challenge in separating benign, non-sensitive code from sensitive, business-critical code, as both may have structural similarities, but business logic could be very different, and the current security systems are not able to handle these variations. In this paper, we propose a novel solution, Source Code Guardrail (SCG), powered by an AI classification model, that automatically categorizes source code as either sensitive (custom, production-grade, or confidential) or non-sensitive (dummy or generic code). Our solution leverages source code embedding models (unixcoder) to convert the code into a language-agnostic numerical vector representation that captures its full details, including semantic meaning, structure, and functionality. These numerical embedding vectors were used as the input feature of the dense layer classification model. In this work, the classification model and UnixCoder are trained together as a merged network so that the model can also learn the code embeddings based on classification loss. The proposed model was trained on manually annotated 8000+ source codes of different languages, which were taken from multiple sources, ensuring a diverse and representative dataset. The proposed model achieves 91.19
Ring signatures allow a signer to sign a message for a set of signers, called a ring, ensuring that the message is signed by one of the signers in the ring, while not revealing which one actually signed it. Constructing ring signature schemes that are tightly secure and achieve signature sizes that grow logarithmically with respect to the ring size remains a fundamental theoretical challenge. To our knowledge, three prior works by Libert et al. (ESORICS 2018), Tang (ICICS 2021), and Hara and Tanaka (Theor. Comput. Sci. 2021) proposed tightly secure ring signature schemes with logarithmic signature size. However, these schemes rely on the hardness of decisional problems. In this paper, we propose a new tightly secure and logarithmic-size ring signature scheme. Our scheme is secure under the discrete logarithm (DL) problem over pairing-free groups in the non-programmable random oracle model (NPROM). Our construction is based on the -protocol of Groth and Kohlweiss (EUROCRYPT 2015), instantiated with our modified Pedersen commitment scheme. Then, we derive our ring signature scheme by applying the randomized Fischlin transformation by Kondi and Shelat (ASIACRYPT 2022) to the above -protocol. The security analysis is obtained by modifying the analysis of the randomized Fischlin transformation by Hashimoto et al. (ePrint Arch. 2024).
In this work, we explore the possibility of unconditionally secure universally composable (UC) commitments, a very relevant cryptographic primitive in the context of secure multi-party computation. To this end, we assume the existence of Physically Uncloneable Functions (PUFs), a hardware security assumption that has been proven useful for securely achieving diverse tasks. In prior work [ASIACRYPT 2013, LNCS, vol. 8270, pp. 100–119] it was shown that a protocol for unconditional UC-secure commitments can be constructed even when the PUFs are malicious. Here, we report an attack to this protocol, as well as a few more issues that we identified in its construction. To address them, first we revise some of the previous PUF properties, and introduce new properties and tools that allow us to rigorously develop and present the security proofs. Second, we propose two different ways for making the commitment scheme secure against the attack we found. The first involves considering a new model where the creator of a PUF is notified whenever the PUF is queried and the second involves restricting adversaries to only being able to create stateless malicious PUFs. Finally, we analyze the efficiency of our schemes and show that our constructions are advantageous in this respect compared to the original proposal.
Third-party private set intersection (PSI) enables two parties, each holding a private set to compute their intersection and reveal the result only to an inputless third party. In this paper, we present an efficient round-optimal third-party PSI protocol. Our work is motivated by real-world applications such as contact tracing whereby expedition is essential while concurrently preserving privacy. Our construction only requires 2 communication rounds and attains a near-linear computational complexity of O(n^1+ε) for large dataset size n, where ε >0 is any fixed constant. Our improvements stem from algorithmic changes and the incorporation of new techniques to achieve a tight asymptotic bound. Furthermore, we also present a third-party PSI cardinality protocol which has not been explored in prior third-party PSI work. In a third-party PSI cardinality setting, only the third-party obtains the size of the intersection and nothing else. Our construction to achieve the cardinality functionality attains a quasilinear computational complexity for the third-party.
Searchable symmetric encryption allows clients to outsource their databases to a semi-trusted cloud server while enabling private searches. The Oblivious Cross-Tag (OXT) protocol is a fundamental approach to conjunctive keyword search, ensuring that search performance scales with the least frequent keyword while introducing keyword pair result pattern (KPRP) and intersection result pattern (IP) leakages. However, recent studies show that the KPRP leakage in OXT can be exploited, allowing the cloud server to infer information about the client database. Several works have aimed to mitigate this issue, with Doris being the first non-interactive OXT-based scheme to hide KPRP and IP leakages. However, this comes at the cost of increased storage overhead. In this work, we propose a Doris-based conjunctive SSE scheme with improved storage efficiency. We replace the XOR filter in Doris with our XEBFF filter, which formalizes XOR filters and Binary Fuse Filters. Additionally, we introduce a frequency estimation approach using Count-Min Sketch to efficiently determine the least frequent keyword, which all previous OXT-based schemes overlook. Our scheme reduces storage overhead by 8
Federated Learning (FL), as a distributed model training paradigm, has garnered significant attention and practical application. Recently, increasing research efforts have focused on Personalized Federated Learning (PFL) as an effective solution to address data heterogeneity in FL systems. However, existing studies reveal that PFL remains vulnerable to stealthy yet harmful backdoor attacks. Furthermore, current federated learning (FL) algorithms designed to defend against backdoor attacks demonstrate significant degradation in model performance and substantial decline in defensive effectiveness when applied to personalized scenarios. To bridge this research gap, we propose a robust PFL framework against backdoor attacks. Our framework incorporates a three-tier defense mechanism: (1) Clients initially purify potential model backdoors through adversarial example generation; (2) An alternating training strategy for hierarchical models is employed to block backdoor attacks while generating personalized head models; (3) The server implements a trimmed aggregation mechanism to mitigate malicious client impacts. Comprehensive experiments on three benchmark datasets demonstrate the framework’s effectiveness, showing superior model performance in personalized scenarios while achieving up to 89
The proposal of registered attribute-based signature(registered ABS) recently eliminates the long-standing key-escrow problem that has plagued classical attribute-based signature(ABS) for a long time. It allows users to generate public and secret key pairs themselves and register their public key and attribute with a key curator. The key curator is fully transparent and retains no secrets. Due to the introduction of public user registration, unlike classical ABS where only attribute authorities are aware of user's attribute information, in registered ABS, all other users in the system can directly view a user's attributes during registration phase. This is not conducive to the deployment of the scheme in scenarios that pursue user privacy. In this paper, we propose a new model of registered ABS that addresses the key-escrow problem while preserving users' attribute privacy. We have provided a complete security definition of our registered ABS and proposed a detailed construction framework. Although our scheme necessitates a trusted authority, it ensures that adversaries cannot forge user signatures even if the authority is compromised.
With the increasing scale and complexity of global maritime traffic, ensuring the safety of autonomous vessel navigation has become a critical challenge. This paper presents a deep reinforcement learning (DRL) approach for autonomous maritime collision avoidance, with a focus on ensuring safety under both nominal and adversarial conditions. A policy is trained using local observations of surrounding vessels to generate COLREGs-compliant maneuvers in decentralized multi-agent scenarios. The method is evaluated in diverse encounter geometries inspired by the Imazu problem set, demonstrating the agent's ability to generalize to unseen head-on, crossing, and overtaking situations. To enhance robustness against positioning interference, we introduce an anomaly detection mechanism based on Inertial Navigation System estimation. During GPS spoofing attacks, the system compares GPS and INS position estimates, and penalizes discrepancies in the reward function, enabling the agent to identify and mitigate spoofed signals without relying on external supervision. Experimental results across multiple scenarios confirm the agent's ability to preserve safe trajectories and avoid collisions, even under sensor-level adversarial attacks.
Constant-time modular inversion ( ) is a critical operation in secure elliptic curve cryptosystems. Existing algorithms include those by Bos, Bernstein and Yang, and Jin and Miyaji, denoted as , , and . While is constant-time, it incurs redundant computations in its iteration function. reduces iteration cost but increases the number of iterations, whereas balances both by incorporating a table-lookup function. We propose two new algorithms, 𝖪𝖬_1 and 𝖪𝖬_2 , that improve upon by reducing table lookups and lowering iteration count. We prove that their iteration count is reduced by two compared to , and we implement both algorithms for practical evaluation. Experiments over NIST prime fields (P192, P224, P256, P384, P512) show that 𝖪𝖬_1 and 𝖪𝖬_2 achieve fewer average clock cycles than existing algorithms. These results demonstrate that the proposed algorithms are efficient and secure choices for modular inversion in side-channel-resistant elliptic curve cryptography.
In recent years, blockchain-based cloud storage systems have seen rapid development. However, existing systems lack a crucial feature: keyword search, primarily due to file encryption. To address this limitation while maintaining privacy, significant research attention has turned to blockchain-based symmetric searchable encryption (SSE) schemes. However, achieving service fairness in the blockchain security model poses challenges, particularly in ensuring security for both users and service providers, while both can be malicious. To address this challenge, this paper introduces SEARCHAIN, a novel blockchain-based SSE system. We introduce a new role in the scenario, verifier, for the blockchain nodes. With the assistance of honest verifiers, service fairness can be guaranteed. We propose a novel committee selection algorithm for the system to select verifier committee members, ensuring their voting powers are proportional to the number of active SSE services they manage. Verifiers are required to perform result verification work for SSE services to participate in mining. With this approach, the committee selection algorithm provides both security and incentives for the system. We implement a prototype of SEARCHAIN and deploy the system to a local test network. The experimental results demonstrate the feasibility and efficiency of SEARCHAIN.
Blockchain has suffered a series of high-impact financial attacks in recent years, underscoring the urgent need for robust security mechanisms. As it becomes increasingly central to native financial infrastructures such as cryptocurrencies and decentralized finance (DeFi), ensuring the financial security of blockchain ecosystems has emerged as a critical challenge. While a growing body of research has focused on smart contract vulnerabilities, relatively limited attention has been paid to the detection of anomalous behaviors at the account level–despite their central role in on-chain financial threats. Existing anomaly detection approaches predominantly rely on static rules or pre-defined patterns, which are inadequate for modeling the dynamic and covert behaviors of malicious accounts. To address this gap, we propose AccountCatcher, an anomaly detection framework based on graph embedding and graph attention network(GAT). By modeling account interactions and capturing abnormal transactional patterns, our approach effectively identifies high-risk accounts. Experiments on real-world Ethereum data demonstrate that AccountCatcher attains 94.4
With the increasing role of Machine Learning (ML) and Deep Learning (DL) in various domains, their application in enhancing Network Intrusion Detection Systems (NIDS) has gained significant attention. Traditional NIDS approaches often rely on correlation-based detection, which may lead to misleading or fake correlations, failing to align with real-world use cases. Addressing this issue requires additional features, new datasets, and the development of new solutions. However, the rapid advancements in ML and DL pose challenges for timely deployment, as training, testing, and evaluating new models against existing solutions can be time-consuming. The large size of real-world datasets also contributes to high computational costs and extended training times, limiting the practical use of ML-based NIDS in dynamic environments. To tackle these challenges, this paper contributes to the field of NIDS in three key aspects: employing Reinforcement Learning (RL) to accelerate and optimize the model tuning process; introducing an efficient data preprocessing pipeline specifically designed for NIDS, which enhances data quality and feature representation; and proposing a novel sampling strategy that determines an optimal dataset size both in terms of total records and class-level balance. By integrating model tuning with the proposed method on dataset sampling, this research uses a smaller sampling size of 3,898 records and achieves a higher F1 score of 93.20, compared to the state-of-the-art statistical sampling method on the same NIDS dataset.
Federated Learning (FL) enables collaborative model training without sharing raw data, but remains vulnerable to poisoning attacks from malicious clients. Existing defenses are often reactive and require costly model retraining, making them inefficient and impractical for real-time protection. We propose FedCleaner, a server-side dual-mechanism framework that combines: Proactive Layer-Wise Anomaly Detection to identify poisoned updates in real time; Retroactive Contribution Erasure to efficiently unlearn malicious client influences without retraining. Experiments on datasets show that FedCleaner provides a scalable, privacy-preserving, and regulation-compliant solution to defend FL systems against persistent poisoning threats.
Tool-using LLM agents for health monitoring raise critical privacy concerns as they share sensitive patient data with cloud providers and third-party models. This study presents HealthAgent, a privacy-preserving LLM agent framework that protects both user queries and multi-modal sensor data through homomorphic encryption. HealthAgent enables an LLM orchestrator to coordinate specialized AI models for complex health assessments while processing all data in encrypted form. The system achieves 95
Statistical Cryptanalysis (SC) and Side Channel Analysis (SCA) are powerful cryptanalytic techniques used to validate their secure implementations and verification. However, the practical certainty of the success of these attacks is probabilistic and typically not guaranteed to provide a full deduction of the cryptosystem due to time and computational complexity restrictions introduced by the design choices of the primitive. Most modern cryptosystems are designed to have an acceptable degree of resistance to SC and SCA. In most cases, SC resistance is established on the standard algorithm, and their success is on reduced versions of the cipher. In addition, the resistance to SCA is at the implementation level, not on the mathematical structure of the algorithm. Similarly, the countermeasures utilized in the cryptosystem's implementations make SCA success rates lower in practice. Therefore, SCA and SC independently might not be powerful enough to always yield practical results. We propose an efficient framework, called the DHABI framework, to combine different classes of theoretical and hardware cryptanalysis techniques, i.e., SC and SCA, making cryptanalytic attacks more practical in the real world. This aims to reduce the security of real cryptosystems by leveraging the strengths of each technique, yielding higher success rates of the attack. In this work, we have shown an application of the framework to the cryptanalysis of a full SPECK-32/64 ARX-based cipher. Independently, we also provide 78 multiple-differential characteristics of 6-round SPECK-32/64 with a combined differential probability of. 2(-9.12).
