Blockchain Papers

Muhammad Ajmal Azad, S A Shah

No abstract is available for this record.

Ayyana Prabu J, Ramkumar M. P, Emil Selvan G S R, Muthumeena S · 6 authors

The pervasive adoption of Electronic Health Records (EHR) has improved the competence of healthcare storage, management of patient data and retrieval. These increase the vulnerability of EHR records through data breaches, unauthorized access, and proliferation of insider threat behaviors, with limited compatibility of health records. To address this issue, blockchain-based systems were introduced; however, they encountered leakage of sensitive information and have limited scalability. Concerning this, a Multi-Layer Blockchain framework integrated with Zero-Knowledge Proof (ZKP) (MLB-ZKP) was proposed to secure the EHR systems by privacy-preserving verification and reduce sensitive data exposure. The multi-layer blockchain method ensures the data storage, secure transaction validation process and efficient access control process. The proposed framework integrates the ZKP mechanism to aid the secure authentication process and data verification process without revealing the sensitive information. The entire framework ensures the patientcentric data ownership, secure cross-institutional data sharing and a governing privacy protection process. To validate the effectiveness of the proposed framework, available benchmark datasets were used. The experimental results show that the proposed framework exhibits increased privacy level, reduced data exposure risk and storage efficiency.

Rohit Kumar, U. Padmavathi, Neha Agrawal

No abstract is available for this record.

Raneem Khaled AlFadhel, Mohammad Ali A. Hammoudeh

With the growing volume of sensitive data stored and processed in cloud environments, conventional security models are no longer sufficient to guarantee privacy, integrity, and trust. This paper proposes a blockchain-based framework that integrates Zero-Knowledge Proofs (ZKPs) and homomorphic encryption (HE) to enable secure and privacy-preserving data sharing. ZKPs are employed to verify user access rights without exposing identities or underlying information, while HE allows computations to be performed directly on encrypted data, ensuring confidentiality is preserved throughout the data lifecycle. The proposed framework addresses the limitations of existing approaches that either lack encrypted computation capabilities or expose sensitive data during processing. Formal and informal analyses demonstrate the feasibility of the model in terms of encryption time, ZKP verification latency, and computation overhead. The framework is designed to be applied initially in the healthcare sector and aligns with national digital transformation initiatives such as Saudi Vision 2030.

Vallamkonda Jyothi, Dama Rishitha Naidu, C N Ravindra Kumar, Dunna Lalith Krishna Kanth · 5 authors

The accelerated digitalization of academic qualifications requires certificate authentication systems with not only the capability not to be tampered with, but also privacy-assuring and scalable. Although certificate verification using blockchain guarantees immutability and transparency, current solutions have high transaction costs, low scalability, and a loosely applied guarantee of privacy. In this research, proposes a new Privacy-Preserving and Scalable Blockchain-Based Certificate Authentication System, which combines Zero-Knowledge (ZK) rollups with an AI-based system of trust scoring. ZK-rollups save a lot of gas through batching certificate transactions and generating succinct cryptographic proofs, which enhances throughput and makes operational costs less. Zero-knowledge proofs can be used to facilitate privacy by providing the selective disclosure so that the verifiers can verify the validity of the certificates without access to sensitive personal information. Also, trust scoring model is an AI-based model that dynamically analyzes the actions of the validators to identify anomalies, collusion, and malicious actions. The experimental analysis shows significant reduction in transaction costs (as much as 80 percent), lower verification latency as well as resilience against Sybil and coordinated attacks. The suggested model provides a privacy-conscious, secure and scalable decentralized certificate authentication model.

V. Hemamalini, Likhith Kumar Reddy Ponnapati, Aviv P Joji, Garv S Rathore

The rapid proliferation of digital educational credentials has intensified challenges related to credential fraud, privacy infringement, and reliance on centralized verification infrastructures. Conventional credential verification mechanisms depend on Public Key Infrastructure managed by centralized Certificate Authorities, resulting in single points of failure, limited scalability, and increased operational overhead. Although blockchain-based credentialing approaches introduce immutability and tamper resistance, such solutions often suffer from a privacy–transparency trade-off, as verification commonly requires exposing complete credential data or associated metadata on public ledgers. A privacy-preserving decentralized framework for Self-Sovereign Identity and Verifiable Credentials is presented to address these limitations. The proposed architecture combines distributed ledger-based trust anchoring with zero-knowledge cryptographic techniques to enable secure and confidential credential verification. Credential commitments are immutably anchored on a high-throughput consensus network, while verification is performed using Zero-Knowledge Succinct Non-Interactive Arguments of Knowledge, enabling proof of credential validity without revealing underlying secrets or sensitive attributes. Verification is conducted without continuous involvement of the issuing authority, thereby eliminating centralized dependency and improving system availability. Experimental evaluation demonstrates that the framework supports sub-second proof generation and millisecond-level verification, while maintaining minimal on-chain storage and predictable operational costs. The results indicate that the proposed approach provides a scalable, efficient, and privacy-respecting solution for decentralized academic credential verification in modern digital identity ecosystems.

Ramesh Kumar Veerapaneni, Radhakrishnan Delhibabu

The rapid aging of the global population necessitates automated healthcare environments, yet current Intelligent Room architectures relying on centralized cloud servers face critical challenges regarding data opacity and single points of failure. This paper proposes a novel architecture that synergizes Distributed Ledger Technology (DLT) with Federated Learning (FL) to create a trustless, immutable audit trail for patient monitoring. Unlike traditional FL approaches, we introduce a blockchain-based aggregation mechanism that eliminates the central authority. Furthermore, to address the resource constraints of edge devices such as smartphones, we implement a specific Lightweight Neural Network (L-CNN) utilizing depthwise separable convolutions. The proposed system ensures that patient data remains local while model updates are cryptographically verified on-chain, offering a scalable, low-cost solution for resource-constrained healthcare environments.

Mia Bodycomb, Ali Anaissi

Swarm Learning (SL) offers a transformative solution to the challenges posed by growing data security regulations and privacy concerns. It creates new opportunities for research in fields such as healthcare, finance, and smart technologies. This decentralized machine learning framework harnesses the collective intelligence of distributed nodes, each holding private data, and uses blockchain technology to ensure data privacy. The framework constructs a shared model by aggregating insights from each node without compromising the security of local data. Motivated by the goals of enhancing model performance and deepening the understanding of model aggregation, this study systematically tested various merging strategies on three datasets—MNIST, BloodMNIST, and Blood Cell Cancer (ALL)—within a simulated Swarm Learning environment. As a result, we developed the Adaptive Performance-Based Merge Strategy (AP-BMS), a novel method that dynamically selects the optimal merging algorithm within the Swarm network based on continuous model evaluations. This strategy improved performance by approximately 1% on the MNIST dataset, 6% on BloodMNIST and 4% on the Blood Cell Cancer (ALL) dataset. The AP-BMS marks a significant advancement in local model aggregation and further accelerates the evolution of Swarm Learning and its application in secure, decentralized machine learning environments.

Idris Adedamola Abdulhameed

Location privacy insparseInternet of Vehicles is difficult to ensure due to limited anonymity, predictable mobility, and prolonged tracking windows. Existing silent-period and pseudonym-based schemes generally assume dense traffic and thus degrade under low-density conditions. This work proposes ZK-V2XChain, a lightweight privacy-preserving framework that integrates a Random Silent Period (RSP) mechanism with blockchain-based Identity Token (IT) authentication and Zero-Knowledge Proof (ZKP) validation. The framework explicitly models sparse-network behavior and enables adaptive, verifiable privacy without compromising efficiency. We design a privacy-preserving IT issuance process using simulated smart contracts and implement decentralized IT verification through RSU–blockchain interaction. Using SUMO mobility traces, ns-3.45 simulations, and MATLAB-based privacy analytics, results show that ZK-V2XChain achieves higher entropy than SAP, RFPM, GLS, and CPS, and approaches the performance of OBS. The maximum anonymity-set size reaches 10.88, and communication overhead remains low (940 bytes). Ablation studies highlight the complementary roles of RSP, ZKP, and blockchain in balancing uncertainty, responsiveness, and issuance stability.

Laila Khalid, Muhammad Usman Akhtar, Muhammad Khalid, Iftikhar Ahmed

The evolving technology in AI and distributed systems requires ethical concepts of how sensitive data can be verified without breach of privacy. Conventional AI systems present the following critical concerns: exposure of data, breach of privacy, and ethical issues concerning transparent but confidential computation. This chapter is a full-fledged cryptographic proof, Zero-Knowledge Proofs (ZKPs), which makes it possible to deploy AI ethically by verifying privacy. The framework is supported by mathematical underpinnings to enable model validation and training verification, as well as federated learning without the underlying datasets or parameters of the models. The chapter shows that ZKPs can be used to meet ethical AI without compromising privacy. It can be used in healthcare, finance, and voting systems where ethical concerns require verification and confidentiality. This chapter offers a new method of dealing with core ethical dilemmas in AI systems and safeguarding privacy and security in algorithmic decision-making exercises.

Dr.M.Swapna Dr.M.Swapna, MORTHALA RACHANA, NERALLA DEEPIKA, VENNU LEELAPRASAD · 5 authors

Healthcare AI systems put a lot of importance on keeping medical data private because it is very sensitive. AI-driven diagnostic models could help doctors make better decisions, but they need a lot of different patient data sets, which are often kept separate from each other at different hospitals. Federated Learning (FL) is a decentralised way to solve this problem by letting multiple people train a model together without sharing data in one place. But conventional FL frameworks continue to encounter challenges related to trust, transparency, and data integrity. This paper puts forth a Blockchain-Enabled Federated Learning Framework to facilitate secure, privacy-preserving, and auditable medical diagnosis across decentralised healthcare systems. This system uses blockchain's unchangeable nature and smart contract features to make sure that model updates can't be changed, contributions can be tracked, and trust between the entities involved is higher. This combination makes AI-driven diagnostics possible without putting patient privacy, regulatory compliance, or institutional integrity at risk.

Malika Abid, Mohammed Kamel Benkaddour, Mohamed Benouis, Yekta Said Can

No abstract is available for this record.

Ammar Ahmed, Amna Saleem Sheikh, Nasir Ayub, Umair Ghafoor · 6 authors

Federated Learning (FL) is an approach that allows numerous users to train a single machine learning model with the oversight of a central server, and with their training data stored locally on their devices. The approach is relevant in alleviating the risks associated with violations in data privacy. It is a process by which a pool of clients collaborates towards solving machine learning problems, with a central coordinator being the one who coordinates the entire process. The paper will review the latest advances in privacy-preserving federated learning and discuss it in the context of machine learning. It assesses privacy-related solutions, which are already in existence, such as secure aggregation, meta-learning, blockchain technology, decentralized training, searchable encryption, and data privacy mechanisms and zero-knowledge proofs. Federated learning (FL) is an emerging technology that can be used in the realm of the intelligence of the Internet of Things. However, the information that is model-related can be shared in FL and reveal the sensitive data of the participants. In this regard, we propose a new privacy-preserving FL framework, which is founded on a new chained secure multiparty computing technique, which we call chain-PPFL. The scheme we are proposing is based mostly on two mechanisms: 1) a single-masking mechanism, which protects the information that is exchanged between participants in a serial chain frame and 2) a chained-communication mechanism, which allows the masked information to be communicated between participants in a serial chain frame. We run large-scale experiments with respect to simulation by comparing the training accuracy and the leak defence to other state-of-the-art schemes with two publicly available data sets (MNIST and CIFAR-100). We established data sample distributions (IID and NonIID), and training models (CNN, MLP and L-BFGS) in our experiments. The experiment results show that the chain-PPFL scheme can offer a realistic privacy preservation (which is the same as the various privacy with ϵ to near zero) to FL at the cost of communication, and without compromising the accuracy and convergence rate of the training model.

Murari Kumar Singh, Sanjeev Kumar Pippal, Vishnu Sharma, Ashish Kumar Chakraverti

No abstract is available for this record.

Yue Wang, Xin Yan

Federated graphs learning for graphs enables multiple clients to share model knowledge and engage in collaborative training while ensuring user data privacy. Nevertheless, federated learning for graphs also faces various security threats, such as privacy leakage and malicious attacks. On the other hand, compared with other security strategies, differential privacy offers low cost and high efficiency in protecting data in federated learning, yet it can compromise the training accuracy of federated learning for graphs and, in some cases, severely degrade training performance. Therefore, this paper considers noise-sensitive scenarios where even a small amount of noise can significantly impact training, and integrates knowledge distillation with distributed differential privacy federated learning for graphs. This approach enhances model training accuracy under noise-sensitive conditions while mitigating the adverse effects of differential privacy noise on training, all while ensuring model security. In addition to leveraging differential privacy to protect data and parameter privacy, we further aim to defend against malicious client attacks. By establishing a global consensus on the gradient clipping range, we use zero-knowledge proofs to provide sampled verification of the gradient range, demonstrating that the parameters uploaded by clients have been correctly clipped during training. Parameters that fail the verification are discarded, thereby further enhancing security.

Prakash Reddy Vanga

Federated learning represents a paradigm shift in distributed machine learning by enabling collaborative model training across decentralized nodes while maintaining data privacy at source locations. It helps bridge the gap between artificial intelligence-driven development guidelines and the regulatory mandates laid down by data protection legislation. A decentralized architecture transmits only the model updates to aggregation servers; this reduces privacy breach exposure and compliance violation risks and also eliminates raw data centralization. Federated learning helps build production-ready systems across healthcare, finance, and edge computing environments, owing to the maturities that have occurred in cloud infrastructure. This is a transition from the erstwhile theoretical frameworks it used to have. Architectural advantages are supplemented by privacy-preserving mechanisms like differential privacy and secure aggregation protocols, which facilitate organizations to leverage collective intelligence without exposing sensitive information. Robust platforms for privacy-critical applications can be synthesized by the integration of cloud-native security services, cryptographic enhancements, and edge computing optimization. Courtesy of emerging solutions that cater to model fairness, communication efficiency, and data heterogeneity, federated learning's practical applicability across diverse organizational contexts and regulatory domains continues to advance.

Rana Hassam Ahmed, Muhammad Sarfraz Khan, Amirmohammad Delshadi, Naseer Ahmad · 5 authors

Internet of Medical Things (IoMT) provides the possibility to conduct continuous monitoring of health, perform intelligent diagnostics, and make a clinical decision based on data. Nonetheless, there are security, privacy, scalability, latency, and energy issues with large-scale deployment. Although Federated learning (FL) provides less exposure to data, and blockchain provides trust, current solutions that combine both blockchain and FL have high consensus overhead, fixed privacy, and adversarial resilience. To handle them, we present an Edge-Intelligent Hierarchical Blockchain-IoMT framework that integrates Hierarchical FL (HFL), Adaptive Differential Privacy (ADP), Lightweight Homomorphic Encryption (LHE), Zero-Knowledge Proof (ZKP) authentication, and an Energy-Aware PoS with Edge Learning (PoS-EL) consensus. Hierarchical aggregation minimizes bottlenecks in communication. ADP minimizes security vs utility. ZKP achieves authentication and PoS-EL minimizes energy consumption. Experiments on real-world data demonstrate 99.21% accuracy of detecting anomalies, 34% decreased latency, 41% decreased energy usage, 52 percent lower blockchain overhead and 97 percent resistance to adversarial attacks, which justifies the framework in real-time, mission-critical IoMT systems.

Kaya Alpturer, Constantine Doumanidis, Aviv Zohar

Peer-discovery protocols within P2P networks are often vulnerable: because creating network identities is essentially free, adversaries can eclipse honest nodes or partition the overlay. This threat is especially acute for blockchains, whose security depends on resilient peer connectivity. We present AetherWeave, a stake-backed peer-discovery protocol that ties network participation to deposited stake, raising the cost of large-scale attacks. We prove that, with high probability, either the honest overlay remains connected or a $(1{-}δ)$-fraction of nodes in every smaller component raise an attack-detection flag -- even against a very powerful adversary. To our knowledge, AetherWeave is the first peer-discovery protocol to simultaneously provide Sybil resistance and privacy: nodes prove they hold valid stake without revealing which deposit they own, and gossiping does not expose peer-table contents. A cryptographic commitment scheme rate-limits discovery requests per round; exceeding the limit yields a publicly verifiable misbehavior proof that triggers on-chain slashing. Beyond deposit and slashing, the protocol requires no on-chain interaction, with per-node communication scaling as $O(s\sqrt{n})$. We validate our design through a mean-field analysis with closed-form convergence bounds, extensive adversarial simulations, and an end-to-end prototype built by forking Prysm, a leading Ethereum consensus client.

Vasiliy Krundyshev, Maxim Kalinin, Oleg Vasiliev

In recent years, the distributed ledger systems (DLS) has become an indispensable approach for creating anonymous payment systems in Smart City infrastructures (in transportation and industrial systems, energy planting and distribution, etc.). Although users in such systems are identified indirectly, but through cryptographic protocols, there is a class of attacks aimed at deanonymizing participants by analyzing transactions and constructing a graph of relationships between addresses. To implement these attacks, intruders use statistical analysis methods, graph theory, and specialized utilities for comparing data from external sources, such as exchanges. An analysis of related works in this domain has shown that compromising cryptographic primitives is not a prerequisite for deanonymizing DLS users; in some cases, analyzing public ledger data and the behavioral characteristics of DLS participants is sufficient. The goal of this research is to preserve privacy and develop protocols that minimize the risks of deanonymizing participants in Smart City ledgers built on the UTXO model. This paper presents a developed framework consisting transaction generator, an analyzer for modeling deanonymization attacks, and protocols designed to protect against such attacks. The experimental study has shown that the use of the CoinJoin protocol significantly complicates the deanonymization task and leads to a decrease in the deanonymization accuracy.

Harish Karthikeyan, Antigoni Polychroniadou

Privacy-preserving aggregation is a cornerstone for AI systems that learn from distributed data without exposing individual records, especially in federated learning and telemetry. Existing two-server protocols (e.g., Prio and successors) set a practical baseline by validating inputs while preventing any single party from learning users' values, but they impose symmetric costs on both servers and communication that scales with the per-client input dimension $L$. Modern learning tasks routinely involve dimensionalities $L$ in the tens to hundreds of millions of model parameters. We present TAPAS, a two-server asymmetric private aggregation scheme that addresses these limitations along four dimensions: (i) no trusted setup or preprocessing, (ii) server-side communication that is independent of $L$ (iii) post-quantum security based solely on standard lattice assumptions (LWE, SIS), and (iv) stronger robustness with identifiable abort and full malicious security for the servers. A key design choice is intentional asymmetry: one server bears the $O(L)$ aggregation and verification work, while the other operates as a lightweight facilitator with computation independent of $L$. This reduces total cost, enables the secondary server to run on commodity hardware, and strengthens the non-collusion assumption of the servers. One of our main contributions is a suite of new and efficient lattice-based zero-knowledge proofs; to our knowledge, we are the first to establish privacy and correctness with identifiable abort in the two-server setting.

M.Anitha, K. Sivaraman

The fast proliferation of edge computing has come up with serious issues of data privacy and trust within the distributed networks. The paper introduces a new hybrid system combining machine learning (ML) and blockchain platforms to provide an improved level of data privacy in edge environments. The presented approach is a hybrid of federated learning and blockchain-based secure consensus, which will allow training the models decentrally without exposing sensitive information. An encryption layer that preserves privacy guarantees the safety of the data transfer between edge nodes, whereas smart contracts handle access control and authentication independently. The hybrid infrastructure uses AI to identify anomalies and use the mitigation of threats based on their adaptability, and blockchain with a ready-to-trace immutable ledger generates transparent data. Through experimentations, it is shown that the proposed framework outperforms conventional edge privacy schemes on privacy protection, latency, and data integrity. The model obtained ~98% data privacy protection. The study adds to the coherent model that provides the connection between security, scalability and efficiency in the privacy-sensitive edge-working applications like IoT, medical, and smart cities.

Abdul Hadi, Krishna Mula, Ahmad Bacha, Sreekanth Muktevi · 6 authors

The growing nature and complexity of the cyber threats within the distributed digital infrastructures require collective intelligence without jeopardizing the privacy of data. Federated Learning (FL) is an up-and-coming model that holds potential in training models in a decentralized way; nonetheless, the existing FL models are susceptible to information leakage as a result of model updates and adversarial inference attacks. To overcome these shortcomings, this paper introduces a Zero-Knowledge Federated Learning (ZK-FL) system to detect cyber threats in a privacy-preserving way so that collaboration in the learning process can be secured without sensitive information about the intermediate models and without exposing sensitive data. In the suggested solution, the zero-knowledge proof (ZKP) mechanisms along with federated optimization are combined to make sure that the participating clients can prove the accuracy of their local model updates, revealing no data features. This cryptographic integrity check deters malicious model poisoning, gradient inversion and unauthorized inference of data, improving the confidence of heterogeneous and untrusted parties. Another approach used is a secure aggregation protocol which protects model parameters in the transmission process to guarantee end-to-end confidentiality and integrity. The framework is tested with actual datasets of cyber threat in a distributed environment and adversarial environment. Empirical studies show that the suggested ZK-FL model can be used to obtain a high detection accuracy and robustness on par with centralized learning, and substantially increase privacy guarantees and anti-inference attack. Furthermore, the communication and computation cost that is entailed by zero-knowledge verification is within manageable limits, and thus the solution is feasible to large-scale cyber defence systems. The suggested ZK-FL architecture provides a secure and trusted platform to cooperative cyber threat intelligence, which is a scalable service in privacy-sensitive environments like enterprise networks, critical infrastructures, and edge-cloud security systems.

Ankit Kumar, Andres J. Aparcana-Tasayco, Minjung Kim, David Camacho · 5 authors

The expansion of Internet of Things (IoT) devices brings challenges of data security and privacy preservation in critical infrastructure. The proposed system combines blockchain technology with federated learning (FL) to secure IoT communications. It ensures decentralized model training with immutable and verifiable blockchain records. The study incorporates a lightweight FL model with a two-stage multitask head for binary and multiclass attack detection, enabling efficient deployment in constrained IoT. Federated learning effectively resolves privacy issues by facilitating cooperative model training across dispersed IoT nodes without revealing raw data. It guarantees collaboration based on a trust-aware mechanism that evaluates the reliability of clients and guides the robust aggregation. Blockchain ensures tamper-evident auditing of model updates and supports Byzantine-fault-tolerant (BFT) and Delegated proof-of-stake (DPoS) trust guarantees. Blockchain records cryptographic hashes of model modifications in a tamper-proof ledger, ensuring the legitimacy of the training process. Smart contracts enable the tamper-evident logging of model hashes, and global model convergence is ensured by federated averaging. Experimental tests demonstrate a secure and verifiable collaborative learning enabling model integrity in IoT networks. The study achieved a fast block generation time of 77.3ms, satisfactory model performance with 98% of training accuracy, and 0.992 F1-score alongside meaningful evolution of client trust values. The final testing accuracy of the FL model for the binary class detection is 98.1%. In a multi-class attack scenario, the FL model achieves a strong multi-class attack detection rate for dominant attack types.

Abdullah Melhem, Ahmed Aleroud, Abdullah Al-Mamun, Mohamed I. Ibrahem · 5 authors

The use of web-enabled healthcare analytics has broadened access to machine learning (ML)- and AI-driven cloud models, but it has also created privacy and security challenges. Federated learning (FL) has been used to address data privacy issues; however, deployments of current FL architectures rely on centralized aggregation approaches, thereby creating a single point of failure (SPoF), as a successful adversarial attack on the global model during training or inference can compromise the entire system. These approaches also assume homogeneous data distributions across clients and overlook the constraints and diversity of web-based analytics. To address those limitations, traditional blockchain-based FL systems incorporated distributed ledgers to record model updates and artifacts. However, using the chain as a data ledger to record model artifacts and logs increases consensus overhead and coordination costs. This paper introduces Blockchain-based Clustered Federated Learning (BCFL), an architecture-diverse and cluster-based FL framework. Our approach is coordinated by a lightweight permissioned ledger that eliminates the trusted central aggregator while preserving utility, robustness, and verifiable provenance in web-based healthcare analytics. BCFL records compact provenance metadata on-chain while keeping model parameters off-chain. In addition, by distributing trust across clusters, the design reduces the transfer of adversarial attacks across models by limiting the impact of malicious updates during training and improving reliability at inference time. Experiments on real-world healthcare data and other benchmarks show that BCFL improves the performance of trained AI/ML models and reduces attack success rates compared with several FL baselines.