Zeinab Alipanahloo, Michael Duchesne, Kaiwen Zhang
RzkFL is an end-to-end, privacy-preserving machine-learning framework that fuses Federated Learning (FL) with recursive zero-knowledge proofs (ZKPs) to protect data, models, and users while unlocking verifiable inference. Models are trained entirely on local devices, so sensitive data never leave the premises. The resulting model can be monetized by offering verifiable predictions on a pay-per-use basis. During inference, each customer independently computes predictions using private data, making it essential to verify that these inference results are computed correctly and honestly. Unlike existing approaches that rely on heavy communication or centralized trust assumptions, RzkFL allows each customer to generate a cryptographic proof of correct local inference, which can be succinctly verified without revealing input data or model parameters either by the customer or a third party. The core innovation lies in the use of recursive ZKPs, enabling each customer to generate small, composable proofs for intermediate layers of neural network inference. These proofs are then recursively aggregated into a single succinct proof using the Nova proof folding scheme. Nova’s design eliminates the traditional sequential dependency of recursive proofs by enabling incrementally verifiable computation through a folding scheme. RzkFL supports on-chain verification via Ethereum smart contracts, allowing AI results to flow directly into financial workflows. A decentralized file storage system maintains the integrity and availability of the global model. We introduce specialized circuits for input, hidden, and output layers to optimize proof generation time and gas costs. The customer can generate proof for the entire inference computation or delegate the proof generation for the intermediate layers and the output layer to another party. The design suits privacy-preserving machine learning scenarios where customer devices are resource-constrained. Our results show that RzkFL can significantly reduce proof size and verification costs while maintaining privacy, integrity, and scalability in federated inference. This makes it a compelling approach for real-world decentralized AI systems requiring strong verifiability guarantees.
Ghislain Nkamdjin Njike, Anh-Tu Hoang, Stefan Schulte
Distributed learning (DL) is gaining popularity as it enables clients (e.g., AI Agents) to enhance their machine learning (ML) models’ performance by exchanging knowledge without revealing private datasets. State-of-the-art DL approaches primarily focus on transferring knowledge between heterogeneous clients with diverse model architectures, connecting clients with those that can improve their models, and protecting data privacy. However, they overlook the threat of malicious clients that potentially downgrade the models’ performance by sharing inaccurate knowledge or excluding high-performing clients from the training procedure.Therefore, we introduce the Zero-Knowledge Blockchain-Based Knowledge Distillation Learning Framework (zkBKD). In zkBKD, heterogeneous clients communicate with a blockchain network to discover high-performing clients, verify zero-knowledge proofs (ZKPs) to ensure the correctness of the knowledge shared from other clients, and vote to eliminate malicious clients. We analyze security and privacy risks and show that zkBKD prevents membership, poisoning, and collusion attacks. We conduct extensive experiments on two standard datasets across heterogeneous clients with four model architectures. The experimental results demonstrate that zkBKD relatively improves the average model accuracy of all clients by 25.71%. Even lightweight models such as ResNet-2 achieve up to a 103.35% accuracy gain compared to independent training.
Privacy-Preserving Record Linkage (PPRL) integrates sensitive datasets from independent parties without exposing personal identifiers. Although secure multi-party computation (SMC) and homomorphic encryption ensure strong privacy, they suffer from high computational costs and poor scalability. Encoding-based methods, such as Bloom filters, are lightweight but face quality issues at scale owing to saturation and blocking inefficiencies. This study proposes a scalable, modular PPRL framework for distributed platforms. It combines Bloom filter encoding, Hamming-based locality-sensitive hashing (LSH), and Dice similarity within a MapReduce pipeline on a Hadoop distributed file system (HDFS). The system supports decentralized end-to-end linkage under semi-honest or covert adversarial models. Experiments on datasets with$100,000-500,000$records show linear scalability,$7.2 \times$speedup over cryptographic baselines, and recall degradation linked to filter saturation. A regression model captures the execution-candidate volume relationship, thereby aiding system tuning. The framework supports high-throughput, regulation-compliant linkages for healthcare, finance, and public sector use.
Sheikh Sharfuddin Mim, Doina Logofătu, Gabriel Guerrero-Contreras, Inmaculada Medina‐Bulo
Federated learning (FL) offers a compelling solution to the privacy and compliance challenges that plague the financial industry by enabling decentralized machine learning without the need for raw data sharing. As regulations like GDPR and the GLBA enforce strict data protection requirements, financial institutions are increasingly exploring FL as an avenue for collaborative intelligence. This paper presents a comprehensive analysis of state-of-the-art FL algorithms tailored for finance, evaluates their performance across realistic tasks such as credit scoring, fraud detection, and customer segmentation, and identifies the trade-offs among performance, fairness, privacy, and communication cost. We benchmark twelve prominent FL algorithms, highlight their privacy implications with differential privacy and secure aggregation, and provide practical, strategic recommendations for deploying FL systems in financial environments. The study closes with a discussion on regulatory alignment, deployment challenges, and future directions for ethical and robust FL adoption in finance.
The rapid expansion of multi-cloud ecosystems has intensified the demand for privacy-preserving analytics across untrusted infrastructures. This paper proposes Federated Zero-Trust Analytics (FZTA), a framework that integrates federated learning, zero-trust security, and privacy-enhancing computation to enable secure data collaboration without centralized trust. The design combines continuous identity verification, decentralized policy enforcement, and hybrid cryptography based on homomorphic encryption and differential privacy. Evaluation across three commercial clouds demonstrates that FZTA achieves near baseline model accuracy (within 2% of centralized training) while maintaining (ε<1.2, δ=10−5) differential privacy guarantees and less than 20% computational overhead. The framework resists eavesdropping, replay, and model inversion attacks while meeting compliance standards such as GDPR and HIPAA. Results confirm that strong privacy and federated scalability can coexist under zero-trust conditions, establishing a foundation for secure cross-domain analytics in healthcare, finance, and IoT applications.
Internet of Things (IoT) is transforming traditional agriculture into a more efficient, sustainable, and data-driven industry. By connecting various devices and sensors across the farm, IoT enables real-time monitoring, control, and optimization of agricultural processes. However, there are many security issues to deal with. IoT devices in precision farming collect sensitive data such as soil moisture, nutrient levels, and livestock health information. Unauthorized access to farm data can result in data theft, manipulation, or misuse. This can compromise the integrity of farming operations and potentially harm the environment. In this paper, we explore the mathematical foundations and practical implementations of model aggregation in federated learning (FL), with a particular focus on integration with distributed ledger technologies (DLT). We present a comprehensive analysis of aggregation algorithms, their convergence properties, and security guarantees. Additionally, we survey existing tools and platforms that facilitate federated learning deployments and examine how blockchain technology can address key challenges in federated learning systems including trust, incentive mechanisms, and auditability. Our analysis demonstrates that the combination of federated learning with blockchain creates a robust, transparent, and decentralized machine learning systems suitable for privacy-sensitive applications across precision farming, healthcare etc.
Denis Wapukha Walumbe, Gabriel Ndugu Kamau, Jane Wanjiru Njuki
Proof of Stake (PoS) models are energy-efficient and require limited computational power. These features are critical in telemedicine environments, where resource-constrained devices must handle sensitive data securely. The growing need for auditable and privacy-preserving data storage in telemedicine underscores the importance of PoS models optimized for lightweight devices while complying with strict regulatory requirements, such as the Health Insurance Portability and Accountability Act (HIPAA).This study was guided by two research questions: (i) Which PoS models are lightweight and suitable for telemedicine? and (ii) What features make lightweight PoS models effective for privacy and efficiency in telemedicine? To address these questions, a systematic literature review (SLR) guided by the PICOC framework was conducted to investigate lightweight PoS models that can enhance privacy in telemedicine systems. Out of 2,394 papers studies screened, 55 were included in the analysis. The findings identified Algorand, Ouroboros Praos, Tendermint, Nxt, and Casper CBC as promising candidates. Key enabling features included lightweight voting mechanisms, such as Byzantine Agreement protocols and Verifiable Random Functions, as well as cryptographic techniques like symmetric encryption and multiparty computation. Performance metrics evaluated included latency, throughput, energy efficiency, and battery consumption, with Grey Relational Analysis ranking Algorand highest due to its low latency, high throughput, and minimal energy consumption.
Ahmad A Alsharidah, Devki Nandan Jha, Ellis Solaiman, Bo Wei · 6 authors
Federated learning is a promising approach that enables collaborative machine learning (ML) in distributed environments, such as the Internet of Medical Things (IoMT) while preserving consumer privacy. It allows multiple consumers to collaboratively train a model using their own data, sharing only the locally trained model rather than the raw data. Most existing federated learning systems assume a high level of trust in participating nodes, which is unrealistic in real-world consumer-centric scenarios. Involving untrusted nodes can compromise the integrity of the training process and result in potential data breaches. To address these challenges, this paper presents REWARDCHAIN, a novel federated learning framework that leverages blockchain technology to ensure trust and accountability among untrusted IoMT consumers. By recording all model updates and client contributions on an immutable blockchain ledger, REWARDCHAIN allows auditing of the entire training process and attributing any malicious behaviour to specific nodes. Moreover, we design an incentive mechanism that evaluates contributions based on data quality and participant reputation. This system motivates participants to contribute high-quality data through a reputation-constrained reward allocation. Our evaluations show that REWARDCHAIN effectively balances trust, security, and model performance, facilitating a more secure and effective federated learning ecosystem.
Harsha Kumar A, Preetham Venkatram C, N. Saran, David Daniel · 5 authors
Traditional Electronic Health Record (EHR) systems suffer from critical vulnerabilities in security, interoperability, and patient data control. This paper introduces PolyMed, a novel decentralized platform designed to address these challenges. PolyMed combines blockchain, Artificial Intelligence (AI), and edge computing into a synergistic architecture. It uses the Polygon blockchain for immutable record-keeping and a Decentralized Autonomous Organization (DAO) for transparent governance. Patient identity is secured through privacy-preserving zero-knowledge proofs (ZKPs) and anchored to non-transferable Soulbound Tokens (SBTs), granting users true sovereignty over their data. The platform also includes a Decentralized Finance (DeFi) module to improve healthcare accessibility. Empirical evaluations on the Polygon Mainnet confirm the system's viability, showing sub-4-second transaction latencies and over 90% cost savings compared to legacy systems. The integrated AI model, leveraging a LightGBM classifier on a rich set of engineered features, achieves an Area Under the Curve (AUC) of 0.8543 and an accuracy of 80.33% in emergency detection, demonstrating high reliability on a clinically relevant and imbalanced dataset. By aligning with global standards like General Data Protection Regulation (GDPR) and Health Insurance Portability and Accountability Act (HIPAA), PolyMed offers an integrated platform for patient-centric digital health management.
Privacy-preserving technologies have introduced a paradigm shift that allows for realizable secure computing in real-world systems. The significant barrier to the practical adoption of these primitives is the significant computational and communication overhead that is incurred when applied at scale. In this paper, we present an overview of our efforts to bridge the gap between this overhead and practicality for privacy-preserving learning systems using multi-party computation (MPC), zero-knowledge proofs (ZKPs), and fully homomorphic encryption (FHE). Through meticulous hardware/software/algorithm co-design, we show progress towards enabling LLM-scale applications in privacy-preserving settings. We show the efficacy of our solutions in several contexts, including DNN IP ownership, ethical LLM usage enforcement, and transformer inference.
In edge computing, fusing privacy-sensitive heterogeneous sensor data poses challenges in balancing utility, privacy, and efficiency. Existing approaches like zkFL and zkGPT fall short in end-to-end verifiable LLM-driven fusion for non-IID data. We propose a framework embedding ZKPs into adaptive LLM layers for secure multimodal fusion with provable privacy. Key contributions: (1) context-aware attention for LLM fusion; (2) custom zk-SNARK circuits for full verification; (3) dynamic edge optimizations reducing latency by $\mathbf{2 5} \boldsymbol{\%}$. Theoretical analyses provide -DP bounds and convergence guarantees. Experiments on UCI HAR and CIFAR extensions show 91.8% accuracy, MI-AUC of 0.52, and $\mathbf{4 5 ~ m s}$ latency on Jetson Nano, outperforming zkFL by 3.5% in accuracy and 25% in efficiency.
Zoey Ziyi Li, Hui Cui, Alven C. Y. Leung, Dennis Liu · 7 authors
The increasing incidents of data breaches and personal data misuse highlight the urgent need for robust identity management systems. Self-Sovereign Identity (SSI) emerges as the future solution for digital identity management, underpinned by anonymous credentials (AC) and the distributed ledger technology (DLT) for security measures. However, current SSI models only achieve partial decentralization and none of them can fully meet the complex security requirements of real-world applications. In this paper, we address these limitations by constructing a Decentralized Anonymous Credential (DAC) scheme inspired by large universe attribute-based cryptographic primitives. Building on this foundation, we design a distributed identity management system (DisIMS), a comprehensive SSI system built on blockchain, achieving attribute flexibility, anonymity, unlinkability, revocability and selective disclosure. Compared to earlier blockchain-based identity management systems, our DisIMS allows users to selectively link previous transactions to generate verifiable eligibility proofs for the current transaction without leaking their real identities. We also implement DisIMS on both permissioned (Hyperledger Fabric v2.5) and permissionless (Ethereum Sepolia testnet) blockchains. Experimental results show that batch verification outperforms single verification by reducing execution times by approximately 76% to 81% on Hyperledger Fabric and 50% to 71% on Ethereum, based on 200 tests with 10 to 50 credentials containing 50 attributes each, which demonstrates DisIMS practicality for real-world batch verification scenarios.
Iván Abellán Álvarez, Pol Hölzmer, Johannes Sedlmeir
Digital identity wallets promise significant advancements in digital identity management by offering users a high degree of convenience, security, and control over their data disclosure. However, there is also criticism regarding their privacy guarantees, especially when used in regulated use cases that require high levels of assurance on the correctness and binding of a legal identity. In this paper, we present a comprehensive privacy model and analysis of one of the most prominent digital wallets – the European Digital Identity Wallet (EUDIW) – as specified by the Architecture and Reference Framework (ARF) and the eIDAS 2.0 regulation. We employ a suite of qualitative privacy risk assessment methods to systematically map and evaluate information flows in three key use cases. Our analysis identifies multiple privacy risks – including linkability, identifiability, and excessive attribute data disclosure – and reveals that although the ARF is designed to comply with privacy-by-design principles, inherent design choices, such as the reliance on SD-JWT and mDOC data formats, as well as the concept of a Wallet Unit Attestation (WUA), retain risks to user privacy. Building on our findings, we then highlight how advanced Privacy-Enhancing Technologies (PETs), such as (general-purpose) Zero-Knowledge Proofs (ZKPs), can reduce or mitigate some of these risks.
Energy consumption in Federated Learning (FL) has emerged as a major challenge due to the growing deployment of intelligent edge devices and the increasing complexity of machine learning models. FL enables collaborative model training across decentralized data sources without transferring raw data, thereby reducing communication overhead and enhancing data privacy by design. These features make FL particularly suitable for applications in healthcare, finance, and industrial IoT, where data sensitivity and resource constraints are critical. This paper provides a comprehensive survey of energy-efficient techniques in FL, classifying them into four main categories: model compression (including pruning and quantization), communication optimization, client selection, and hardware-aware strategies. The paper presents a unified taxonomy and discusses the strengths, limitations, and trade-offs of each approach. A comparative evaluation framework is introduced to assess energy savings, model accuracy, communication cost, and deployment feasibility. By analyzing current trends and open challenges, this review offers valuable guidance for researchers and practitioners in the development of scalable, energy-aware, and privacy-preserving federated learning systems.
The convergence of federated learning (FL) and differential privacy (DP) presents a transformative approach to secure predictive modeling in healthcare finance, where safeguarding sensitive patient and financial data is paramount. Traditional centralized machine learning methods often raise significant privacy concerns due to the necessity of aggregating data from multiple institutions. Federated learning mitigates this by enabling decentralized model training across disparate data sources, such as hospitals, insurance firms, and financial institutions, without exposing raw data. However, FL alone remains vulnerable to inference and reconstruction attacks. To enhance security, differential privacy introduces mathematically rigorous noise mechanisms that obfuscate sensitive information while preserving data utility. This paper explores the synergistic integration of DP and FL for building robust, privacy-preserving predictive models tailored to healthcare finance applications, such as fraud detection, insurance risk scoring, billing optimization, and cost forecasting. We discuss the architectural design, privacy-utility trade-offs, and implementation challenges involved, including issues of scalability, model accuracy, regulatory compliance (e.g., HIPAA and GDPR), and communication overhead. Furthermore, real-world use cases and simulation results demonstrate the efficacy of DP-FL frameworks in delivering secure and accurate predictive insights without compromising individual or institutional privacy. The study concludes by highlighting open research directions and recommending best practices for deploying privacy-enhanced federated learning systems in complex, multi-stakeholder healthcare financial ecosystems.
Sabri Barbaria, Abderrazak Jemai, Halil İbrahim Ceylan, Raul Ioan Muntean · 6 authors
Background: Healthcare data interoperability faces significant barriers, including regulatory compliance complexities, institutional trust deficits, and technical integration challenges. Current centralized architectures demonstrate inadequate mechanisms for balancing data accessibility requirements with patient privacy protection, as mandated by HIPAA and GDPR frameworks. Traditional compliance approaches rely on manual policy implementation and periodic auditing, which are insufficient for dynamic, multi-organizational healthcare data-sharing scenarios. Objective: This study develops and proposes a blockchain-based healthcare data management framework that leverages Hyperledger Fabric, IPFS, and the HL7 FHIR standard and incorporates automated regulatory compliance mechanisms via smart contract implementation to meet HIPAA and GDPR requirements. It assesses the theoretical system architecture, security characteristics, and scalability considerations. Methods: We developed a permissioned blockchain architecture that employs smart contracts for privacy policy enforcement and for patient consent management. The proposed system incorporates multiple certification authorities for patients, hospitals, and research facilities. Architectural evaluation uses theoretical modeling and system design analysis to assess a system’s security, compliance, and scalability. Results: The proposed framework demonstrated enhanced security through decentralized control mechanisms and cryptographic protection protocols. Smart contract-based compliance verification can automate routine regulatory tasks while maintaining human oversight in complex scenarios. The architecture supports multi-organizational collaboration with attribute-based access control and comprehensive audit-trail capabilities. Conclusions: Blockchain-based healthcare data-sharing systems provide enhanced security and decentralized control compared with traditional architectures. The proposed framework offers a promising solution for automating regulatory compliance. However, implementation considerations—including organizational readiness, technical complexity, and scalability requirements—must be addressed for practical deployment in healthcare settings.
This article explores the characteristics of data resources from the perspective of production factors, analyzes the demand for trustworthy circulation technology, designs a fusion architecture and related solutions, including multi-party data intersection calculation, distributed machine learning, etc. It also compares performance differences, conducts formal verification, points out the value and limitations of architecture innovation, and looks forward to future opportunities.
Xiangyun Tang, Minyang Li, Tao Zhang, Yijing Lin · 7 authors
In low-altitude networks, various aerial platforms such as unmanned aerial vehicles and airships cooperate to provide services including real-time monitoring, emergency response, and data collection. These platforms often operate with limited computing resources, restricted energy supply, and unstable wireless connectivity, making centralized data processing inefficient and prone to privacy risks. Federated Learning (FL) provides a promising solution by enabling multiple platforms to collaboratively train a shared model without exchanging raw data. However, the presence of Byzantine clients and a potentially malicious server poses serious threats to the robustness and trustworthiness of FL in such environments. Existing Byzantine-robust FL methods typically assume a semi-honest server and rely on auxiliary information such as clean datasets or known attacker ratios, which limits their applicability in dynamic and non-IID scenarios. In this paper, we propose zkFL, a Byzantine-robust FL framework that embeds zero-knowledge proofs to ensure verifiable aggregation under a malicious server. ZkFL allows clients to verify the correctness of server-side aggregation and dynamically adjusts client weights based on inference-guided detection, without relying on external datasets. Each round includes a zk-SNARK proof to guarantee aggregation integrity while preserving gradient privacy. Experiments demonstrate that zkFL exhibits strong robustness and verifiability in both IID and non-IID settings, outperforming prior methods, even in the presence of a malicious server.
With the rapid adoption of diffusion models for visual content generation, proving authorship and protecting copyright have become critical. This challenge is particularly important when model owners keep their models private and may be unwilling or unable to handle authorship issues, making third-party verification essential. A natural solution is to embed watermarks for later verification. However, existing methods require access to model weights and rely on computationally heavy procedures, rendering them impractical and non-scalable. To address these challenges, we propose NoisePrints, a lightweight watermarking scheme that utilizes the random seed used to initialize the diffusion process as a proof of authorship without modifying the generation process. Our key observation is that the initial noise derived from a seed is highly correlated with the generated visual content. By incorporating a hash function into the noise sampling process, we further ensure that recovering a valid seed from the content is infeasible. We also show that sampling an alternative seed that passes verification is infeasible, and demonstrate the robustness of our method under various manipulations. Finally, we show how to use cryptographic zero-knowledge proofs to prove ownership without revealing the seed. By keeping the seed secret, we increase the difficulty of watermark removal. In our experiments, we validate NoisePrints on multiple state-of-the-art diffusion models for images and videos, demonstrating efficient verification using only the seed and output, without requiring access to model weights.
Venkata Sivakumar Musam, Nagendra Kumar Musham, C Siva, S. Karimulla Basha · 6 authors
The field of healthcare monitoring has been revolutionized by a combination of cloud computing and IoT-enabled sensor networks that enable real-time data collection, storage and processing. However, because healthcare data is sensitive, strong security and privacy-preserving measures are required to stop unwanted access and preserve data integrity in multi-authority settings. To propose a collaborative and privacy-preserving sensor cloud architecture for safe, scalable, and fine-grained access control for healthcare monitoring systems by utilizing Zero-Knowledge Proofs (ZKP) and Multi-Authority Attribute-Based Encryption (MA-ABE). The proposed method combines ZKP for authentication and MA-ABE for attribute-based encryption to protect sensitive data. It uses gateways to securely aggregate and transmit data to the cloud, and elliptic curve cryptography to optimize performance. The framework decreased computational overhead by 40% while achieving notable gains in encryption (120 ms) and decryption (150 ms). Outperforming current approaches in healthcare data security and access management, classification accuracy reached 96.5% with improved privacy preservation (96%) and scalability (94%). The suggested system addresses privacy, scalability, and computational efficiency while incorporating cutting-edge cryptographic approaches to provide secure healthcare monitoring. Large-scale, real-time healthcare applications can benefit greatly from its strong solution.
Cross chain interoperability in blockchain systems exposes a fundamental tension between user privacy and regulatory accountability. Existing solutions enforce an all or nothing choice between full anonymity and mandatory identity disclosure, which limits adoption in regulated financial settings. We present VeilAudit, a cross chain auditing framework that introduces Auditor Only Linkability, which allows auditors to link transaction behaviors that originate from the same anonymous entity without learning its identity. VeilAudit achieves this with a user generated Linkable Audit Tag that embeds a zero knowledge proof to attest to its validity without exposing the user master wallet address, and with a special ciphertext that only designated auditors can test for linkage. To balance privacy and compliance, VeilAudit also supports threshold gated identity revelation under due process. VeilAudit further provides a mechanism for building reputation in pseudonymous environments, which enables applications such as cross chain credit scoring based on verifiable behavioral history. We formalize the security guarantees and develop a prototype that spans multiple EVM chains. Our evaluation shows that the framework is practical for today multichain environments.
The verifiability of machine learning models and the privacy of training data have become critical concerns due to their widespread deployment in sensitive applications. Ensuring that a model performs as claimed, without revealing private data or algorithms, is a significant challenge. Zero-Knowledge Proof systems (ZKPs) have emerged as a promising cryptographic solution, enabling the verification of statements without disclosing underlying information. Their integration with blockchain technology further enhances trust and decentralization, offering robust solutions for secure and transparent AI systems. This paper explores the use of ZKPs in machine learning, focusing on privacy-preservation techniques, model verifiability, and confidential AI. It compares the differences and challenges of employing ZKPs in machine learning versus blockchains, highlighting their unique requirements and overlapping benefits. We review the basic concepts of ZKPs, advances such as zkSNARKs and zk-STARKs, and their applications in blockchainbased AI frameworks to ensure data integrity, immutability, and scalability. Furthermore, the paper delves into the practical implications of using ZKPs in AI, providing case studies and analyzing their scalability, performance, and limitations. We conclude by identifying key challenges and presenting future research directions to extend the applicability of ZKPs in AI, particularly in federated learning, model fairness, and decentralized AI pipelines.