Blockchain Papers

Mayuri Kulkarni, Ashish Awate, Makarand Shahade, Bhushan Nandwalkar

No abstract is available for this record.

Weiguang Ning, Yingjuan Zhu, Caixia Song, Hongxia Li · 10 authors

Federated learning, as a novel distributed machine learning mode, enables the training of machine learning models on multiple devices while ensuring data privacy. However, the existence of single-point-of-failure bottlenecks, malicious threats, scalability of federated learning implementation, and lack of incentive mechanisms have seriously hindered the development of federated learning technology. In recent years, as a distributed ledger, blockchain has the characteristics of decentralization, tamper-proof, transparency, security, etc., which can solve the issues encountered in the above-mentioned federated learning. Particularly, the integration of federated learning and blockchain leads to a new paradigm, called blockchain-based federated learning (BFL), which has been successfully applied in many application scenarios. This paper aims to provide a comprehensive review of recent efforts on blockchain-based federated learning. More concretely, we propose and design a taxonomy of blockchain-based federated learning models, along with providing a comprehensive summary of the state of the art. Various applications of federated learning based on blockchain are introduced. Finally, we expand on current trends and provide new perspectives pertaining to this new and exciting development in the field.

Fedwa Mansouri, Mounira Tarhouni, Bechir Alaya, Salah Zidi

No abstract is available for this record.

Andreas Künz, Rodrigo Asensio-Garriga, Jesús García-Rodríguez, Jorge Bernal Bernabé · 7 authors

This paper provides an overview of the integration of privacy preserving Attribute Based Credentials (p-ABC) into the 3GPP 5G Advanced system as well as into the European Project RIGOUROUS (secuRe desIGn and deplOyment of trUsthwoRthy cOntinUum computing 6G Services) for 6G. The differences in the access and device (or user) authentication and authorization in the 3GPP system and the concepts of p-ABC require architectural changes in the main procedures. The novel architecture and procedures of p-ABC support in 5G Advanced are described along with an overview of the 6G RIGOUROUS architecture highlighting the updated zero-touch onboarding of devices into the system. This integration enhances the 5G+ system to support privacy goals such as minimal disclosure or controlled linkability through zero-knowledge proofs.

Nikhil Vanjani, Pratik Soni, Sri Aravinda Krishnan Thyagarajan

In scenarios where a seller holds sensitive data $x$, like patient records, and a buyer seeks to obtain an evaluation of a function $f$ on $x$, solutions in trustless environments like blockchain fall into two categories: (1) Smart contract-powered solutions and (2) cryptographic solutions using tools such as adaptor signatures. The former offers atomic transactions where the buyer learns $f(x)$ upon payment. However, this approach is inefficient, costly, lacks privacy for the seller's data, and is incompatible with blockchains such as bitcoin. In contrast, the adaptor signature-based approach addresses all of the above issues but comes with an "all-or-nothing" guarantee, where the buyer fully extracts $x$ and does not support extracting $f(x)$. In this work, we bridge the gap between these approaches, developing a solution that enables fair functional sales while offering all the above properties like adaptor signatures. Towards this, we propose functional adaptor signatures (FAS), a novel cryptographic primitive and show how it can be used to enable functional sales. We formalize the security properties of FAS, among which is a new notion called witness privacy to capture seller's privacy, which ensures the buyer does not learn anything beyond $f(x)$. We present multiple variants of witness privacy, namely, witness hiding, witness indistinguishability, and zero-knowledge. We introduce two efficient constructions of FAS supporting linear functions based on groups of prime-order and lattices, that satisfy the strongest notion of witness privacy. A central conceptual contribution of our work lies in revealing a surprising connection between functional encryption and adaptor signatures. We implement our FAS construction for Schnorr signatures and show that for reasonably sized seller witnesses, all operations are quite efficient even for commodity hardware.

Yazan Otoum, C. Hu, Eyad Haj Said, Amiya Nayak

Federated learning offers a framework for developing local models across institutions while safeguarding sensitive data. This paper introduces a novel approach for heart disease prediction using the TabNet model, which combines the strengths of tree-based models and deep neural networks. Our study utilizes the Comprehensive Heart Disease and UCI Heart Disease datasets, leveraging TabNet’s architecture to enhance data handling in federated environments. Horizontal federated learning was implemented using the federated averaging algorithm to securely aggregate model updates across participants. Blockchain technology was integrated to enhance transparency and accountability, with smart contracts automating governance. The experimental results demonstrate that TabNet achieved the highest balanced metrics score of 1.594 after 50 epochs, with an accuracy of 0.822 and an epsilon value of 6.855, effectively balancing privacy and performance. The model also demonstrated strong accuracy with only 10 iterations on aggregated data, highlighting the benefits of multi-source data integration. This work presents a scalable, privacy-preserving solution for heart disease prediction, combining TabNet and blockchain to address key healthcare challenges while ensuring data integrity.

Narayana Gaddam

As machine learning spreads into fields of use that demand secure and private authentication, ensuring such authentication is becoming increasingly critical.Zero Knowledge Proofs (ZKPs) have been presented as a cryptographic technique of transforming authentication without data leakage [1].In this research, the use of ZKPs in the AI authentication frameworks is looking into privacy, security and scalability.The model predictions are verified by the proposed system using advanced ZKP protocols like zkSNARKs and zkSTARKs without revealing model parameters or user inputs [3].Our system is able to reach better computational efficiency and lower computation overhead through incorporation of Mystique conversion protocols [7] and fast ZK inference protocols such as ezDPS [6].Results of experiments [5] show that frameworks with ZKP integrated authentication perform better than the standard encryption with respect to both security and performance in decentralized machine learning regimes.Moreover, the solution facilitates verifiability in Federated Learning by integrating blockchain, which helps to increase transparency and trust [4].To overcome the data leakage issue, ZKPs are explored for use in decentralized AI frameworks where secure model deployment is required to generate personalized advice [4].As this research shows, ZKPs offer transformative properties which can be used for authentication in AI systemssuch as in healthcare, finance or IoT network and thus increase the trust in AI driven solutions.

Lorenzo Petrosino, Luigi Masi, Federico D’Antoni, Mario Merone · 5 authors

No abstract is available for this record.

Chenhao Ying, Fuyuan Xia, David S. L. Wei, Xinchun Yu · 11 authors

Harnessing the benefits of blockchain, such as decentralization, immutability, and transparency, to bolster the credibility and security attributes of federated learning (FL) has garnered increasing attention. However, blockchain-enabled FL (BFL) still faces several challenges. The primary and most significant issue arises from its essential but slow validation procedure, which selects high-quality local models by recruiting distributed validators. The second issue stems from its incentive mechanism under the transparent nature of blockchain, increasing the risk of privacy breaches regarding workers’ cost information. The final challenge involves data eavesdropping from shared local models. To address these significant obstacles, this paper proposes a Blockchain-enabled Incentivized and Secure Federated Learning (BIT-FL) framework. BIT-FL leverages a novel loop-based sharded consensus algorithm to accelerate the validation procedure, ensuring the same security as non-sharded consensus protocols. It consistently outputs the correct local model selection when the fraction of adversaries among validators is less than$1/2$with synchronous communication. Furthermore, BIT-FL integrates a randomized incentive procedure, attracting more participants while guaranteeing the privacy of their cost information through meticulous worker selection probability design. Finally, by adding artificial Gaussian noise to local models, it ensures the privacy of trainers’ local models. With the careful design of Gaussian noise, the excess empirical risk of BIT-FL is upper-bounded by$\mathcal {O}(\frac{\ln n_{\min}}{ n_{\min}^{3/2}}+\frac{\ln n}{n})$, where$n$represents the size of the union dataset, and$n_{{\min}}$represents the size of the smallest dataset. Our extensive experiments demonstrate that BIT-FL exhibits efficiency, robustness, and high accuracy for both classification and regression tasks.

Chien‐Ming Chen, Yiru Hao, Saru Kumari, Mohammed Amoon

Consumer electronics play a crucial role in the automotive sector by providing the essential hardware infrastructure necessary for developing autonomous vehicles. These vehicles significantly enhance driving comfort and accessibility, particularly benefiting specific populations such as the elderly and individuals with disabilities. Researchers utilize federated learning, employing homomorphic encryption and differential privacy techniques; however, these methods can impede efficiency. Traditional centralized federated learning presents a risk of single points of failure. To address this issue, we propose an authentication protocol for federated learning, integrated with blockchain technology, within consumer electronics-supported autonomous driving environments. This protocol enhances vehicle authentication efficiency while prioritizing data security and user privacy. Formal analysis verifies the integrity of the protocol. Experimental results demonstrate that the communication overhead of our protocol is reduced by 40.3% compared to existing similar authentication protocols.

S. Y. Xiao, Xiaoge Huang, Xuesong Deng, Bin Cao · 5 authors

To protect user privacy and data security, the integration of Federated Learning (FL) and blockchain has become an emerging research hotspot. However, the limited throughput and high communication complexity of traditional blockchains limit their application in large-scale FL tasks, and the synchronous traditional FL will also reduce the training efficiency. To address these issues, in this paper, we propose a Directed Acyclic Graph (DAG) blockchain-enabled generalized Federated Dropout (FD) learning strategy, which could improve the efficiency of FL while ensuring the model generalization. Specifically, the DAG maintained by multiple edge servers will guarantee the security and traceability of the data, and the Reputation-based Tips Selection Algorithm (RTSA) is proposed to reduce the blockchain consensus delay. Second, the semi-asynchronous training among Intelligent Devices (IDs) is adopted to improve the training efficiency, and a reputation-based FD technology is proposed to prevent overfitting of the model. In addition, a Hybrid Optimal Resource Allocation (HORA) algorithm is introduced to minimize the network delay. Finally, simulation results demonstrate the effectiveness and superiority of the proposed algorithms.

Kasturi Routray, Abhiram Manikonda, Padmalochan Bera

In the age of technological advancement, collaborative e-healthcare emerges as a transformative system eliminating traditional location and accessibility barriers in healthcare services. Here, searchable encryption (SE) plays a key role in enabling healthcare providers to outsource encrypted medical data and search services to third parties like cloud servers, thereby reducing storage and management expenses. This intermediary approach poses challenges of single-point failure, privacy breaches, and potentially untrustworthy results. State-of-the-art public key-based SE methods use a cloud-assisted architecture that doesn’t support reliable and practical searches with fine-grained permissions. Also, such systems require additional support to address potential privacy leakages and ensure data availability at the storage server. To address these concerns, we propose a blockchain-assisted efficient and secure keyword search (BESKS) scheme to enforce fine-grained keyword search privilege control while achieving practical search complexity. Our scheme employs a ciphertext-policy attribute-based keyword search mechanism where keywords are encrypted using expressive access policies to build an inverted index structure. The encrypted indexes are stored on the blockchain while encrypted medical documents are stored on InterPlanetary File System (IPFS) nodes to enhance availability and ensure the reliability and scalability of our approach. Our scheme utilizes blockchain-based smart contract for efficient, secure search operations and ensures financial fairness in fine-grained searches. Search tokens are generated based on user attributes and query keywords to facilitate private searches on-chain. To enhance the search process, our secure index enables exact match for a query keyword in constant time to ensure expensive authorization operations are performed only once. Theoretical analysis suggests that our BESKS is more efficient and secure than state-of-the-art schemes. Prototype implementation results on the Ethereum blockchain network further validate its feasibility for real-world applications, demonstrating the scheme’s practical applicability in collaborative e-healthcare systems.

Mahmoud Ahmad Al‐Khasawneh, Muhammad Faheem, Ala Abdulsalam Alarood, Safa Habibullah · 5 authors

Electronic health records are one of the essential components of health organizations. In recent years, there have been increased concerns about privacy and reputation regarding the storage and use of patient information. In this regard, the information provided as a part of medical and health insurance, for instance, can be viewed as proof of social insurance and governance. Several problems in the past few decades regarding medical information management have threatened patient information privacy. In intelligent healthcare applications, the privacy of patients' data is one of the main concerns. As a result, blockchain is a severe necessity as it can enhance transparency and security in medical applications. Accordingly, this paper uses the design science method to propose a secure blockchain framework for healthcare records management systems. The proposed framework comprises five components: a blockchain network, smart contracts, privacy key management, data encryption, and integration with healthcare information technology. In the proposed framework, healthcare organizations can manage healthcare information securely and privately. Additionally, a secure storage system for electronic records is proposed to meet these organizations' needs. It provides security and privacy for healthcare organizations, especially when managing healthcare information, and also proposes a secure storage system for electronic records to meet the needs of the organizations.

Reza Nourmohammadi, Mohammad Reza Sabramooz, Kaiwen Zhang, Chamseddine Talhi

This paper presents a pioneering federated learning framework that leverages a novel hierarchical aggregation approach, empowering clients to collaboratively generate the global model through multiple levels of aggregation. Additionally, a unique role definition mechanism is integrated into the framework to delineate clients’ roles and tasks in each learning round. Moreover, decentralized storage (e.g. IPFS) and blockchain technologies are employed for storing local models and their corresponding hash pointers, respectively, to improve data availability and integrity. We implemented our solution using using Keras, Scikit-learn, web3 and Solidity. The performance of the proposed framework is evaluated using a genomic breast cancer dataset sourced from the GDC portal, yielding a remarkable 98% accuracy for the global model after 12 rounds of learning while for a centrally trained model on the 70% of the entire dataset the accuracy was 95%. This clearly shows the effectiveness of the proposed framework.

Stefano Avola, Pierpaolo Baglietto, Massimo Maresca, A. Ferretto Parodi

Permissioned Blockchains have been proposed and assessed as a novel and promising approach for improving data sharing, trust, and integration in the context of industrial consortia. One of the reasons that are often considered an obstacle to the adoption of this technology is given by the data privacy and confidentiality issues raised by exposing business transactions details and business relationships in a shared distributed ledger. In this paper we propose a solution based on a specific Hyperledger Fabric (HLF) feature, the Private Collections, which supports data privacy and confidentiality while keeping the advantages that the Blockchain distributed architecture brings.

Guozheng Peng, Xiaoyun Shi, Jun Zhang, Lisha Gao · 7 authors

In the rapidly evolving landscape of Industry 4.0, the complex computational tasks and the associated massive data volumes present substantial opportunities for advancements in machine learning at industry edges. Federated learning (FL), which is a variant of distributed machine learning for edge-cloud computing, presents itself as a persuasive resolution for these industrial edges, with its main objectives being the mitigation of privacy breaches and the resolution of data privacy concerns. However, traditional FL methodologies encounter difficulties in effectively overseeing extensive undertakings in Industry 4.0 as a result of challenges including wireless communications with high latency, substantial heterogeneity, and insufficient security protocols. As a consequence of these obstacles, blockchain technology has garnered acclaim for its secure, decentralized, and transparent data storage functionalities. A novel blockchain-enabled group federated learning (BGFL) framework designed specifically for wireless industrial edges is presented in this paper. By strategically dividing industrial devices into multiple groups, the BGFL framework simultaneously reduces the wireless traffic loads required for convergence and improves the accuracy of collaborative learning. Moreover, to optimize aggregation procedures and reduce communication resource utilization, the BGFL employs a hierarchical aggregation strategy that consists of both local and global aggregation off-chain and on-chain, respectively. The integration of a smart contract mechanism serves to fortify the security framework. The results of comparative experimental analyses demonstrate that the BGFL framework enhances the resilience of the learning framework and effectively reduces wireless communication latency. Thus, it offers a scalable and efficient solution for offloading tasks in edge-cloud computing environments.

Xiaohong Zhang, Xingxing Chen, Shuling Liu, Shaojiang Zhong

In recent years, with the increasing integration of intelligent modules into vehicles, intelligent transportation systems (ITS) have increasingly assumed a pivotal role in augmenting driver safety. As an ITS, vehicle ad hoc networks (VANETs) not only establish a secure traffic environment for users but also provides them with an efficient means of exchanging traffic information. However, during vehicle communication processes, security challenges such as the leakage of vehicle privacy information and tampering with shared task information must be addressed. Therefore, this paper proposes a decentralized anonymous authentication and secure traffic task information-sharing scheme based on blockchain and zk-SNARK. Specifically, the proposed scheme utilizes blockchain technology and the interplanetary file system (IPFS) to securely and efficiently store and share task information in a fully decentralized manner. Vehicle users anonymously participate in tasks within VANETs using pseudonym information generated during registration with a trusted authority (TA). Additionally, by employing zk-SNARK to generate zero-knowledge proofs, the validity and integrity of task information can be verified without revealing any private information. Performance comparisons indicate that the proposed scheme enhances the security of vehicle-to-roadside unit (V2R) and vehicle-to-vehicle (V2V) communications while reducing communication and computational costs.

Hao Lin, Jinze Du

Delegated Proof of Stake (DPoS) is a widely utilized consensus protocol in blockchain-based Internet of Things (IoT) systems. We propose a heuristic algorithm-based accounting rights allocation method, also referred to as the witness election method, which aims to address the challenges in DPoS. The challenges associated with selected witness nodes that do not reflecting majority stakeholder preferences and susceptibility to manipulation of the vote. This method employs the Kendall’s rank correlation as the fitness function to optimize the arrangement of the top-k producers, thereby maximizing stakeholder preferences. We propose a novel heuristic algorithm, termed SP-DEWOA, which combines the differential evolution algorithm and whale optimization with piecewise chaotic mapping to maximize permutation similarity, i.e., stakeholder preferences. To further improve the efficiency of SP-DEWOA, we parallelize SP-DEWOA based on the Spark-based parallelization design. Experimental results demonstrate that the witness nodes selected through SP-DEWOA are consistent with the preferences of the majority of stakeholders. Furthermore, SP-DEWOA has been proven to have high scalability and resilience against vote manipulation.

Sebastian Kolby, Divya Ravi, Sophia Yakoubov

YOSO MPC (Gentry et al., Crypto 2021) is a new MPC framework where each participant can speak at most once. This models an adaptive adversary’s ability to watch the network and corrupt or destroy parties it deems significant based on their communication. By using private channels to anonymous receivers (e.g. by encrypting to a public key whose owner is unknown), the communication complexity of YOSO MPC can scale sublinearly with the total number N of available parties, even when the adversary’s corruption threshold is linear in N (e.g. just under N/2). It was previously an open problem whether YOSO MPC can achieve guaranteed output delivery in a constant number of rounds without relying on trusted setup. In this work, we show that this can indeed be accomplished. We demonstrate three different approaches: the first two (which we call YaOSO and YOSO-GLS) use two and three rounds of communication, respectively. Our third approach (which we call YOSO-LHSS) uses O(d) rounds, where d is the multiplicative depth of the circuit being evaluated; however, it can be used to bootstrap any constant-round YOSO protocol that requires setup, by generating that setup within YOSO-LHSS. Though YOSO-LHSS requires more rounds than our first two approaches, it may be more practical, since the zero knowledge proofs it employs are more efficient to instantiate. As a contribution of independent interest, we introduce a verifiable state propagation UC functionality, which allows parties to send private message which are verifiably derived in the “correct” way (according to the protocol in question) to anonymous receivers. This is a natural functionality to build YOSO protocols on top of.

Carsten Baum, Jens Berlips, W Q Chen, Ivan Damgård · 20 authors

Oblivious Pseudorandom Functions (OPRFs) allow a client to evaluate a pseudorandom function (PRF) on her secret input based on a key that is held by a server. In the process, the client only learns the PRF output but not the key, while the server neither learns the input nor the output of the client. The arguably most popular OPRF is due to Naor, Pinkas and Reingold (Eurocrypt 2009). It is based on an Oblivious Exponentiation by the server, with passive security under the Decisional Diffie-Hellman assumption. In this work, we strengthen the security guarantees of the NPR OPRF by protecting it against active attacks of the server. We have implemented our solution and report on the performance. Our main result is a new batch OPRF protocol which is secure against maliciously corrupted servers, but is essentially as efficient as the semi-honest solution. More precisely, the computation (and communication) overhead is a multiplicative factor <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mi>o</mml:mi> <mml:mo stretchy="false">(</mml:mo> <mml:mn>1</mml:mn> <mml:mo stretchy="false">)</mml:mo> </mml:mrow> </mml:math> as the batch size increases. The obvious solution using zero-knowledge proofs would have a constant factor overhead at best, which can be too expensive for certain deployments. Our protocol relies on a novel version of the DDH problem, which we call the Oblivious Exponentiation Problem (OEP), and we give evidence for its hardness in the Generic Group model. We also present a variant of our maliciously secure protocol that does not rely on the OEP but nevertheless only has overhead <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mi>o</mml:mi> <mml:mo stretchy="false">(</mml:mo> <mml:mn>1</mml:mn> <mml:mo stretchy="false">)</mml:mo> </mml:mrow> </mml:math> over the known semi-honest protocol. Moreover, we show that our techniques can also be used to efficiently protect threshold blind BLS signing and threshold ElGamal decryption against malicious attackers.

Balthazar Bauer, Pooya Farshim, Patrick Harasser, Markulf Kohlweiss

The generic-group model (GGM) and the algebraic-group model (AGM) have been exceptionally successful in proving the security of many classical and modern cryptosystems. These models, however, come with standard-model uninstantiability results, raising the question of whether the schemes analyzed under them can be based on firmer standard-model footing. We formulate the uber-knowledge (UK) assumption, a standard-model assumption that naturally extends the uber-assumption family to knowledge-type problems. We justify the soundness of UK in both the bilinear GGM and the bilinear AGM. Along the way we extend these models to account for hashing into groups, an adversarial capability that is available in many concrete groups—In contrast to standard assumptions, hashing may affect the validity of knowledge assumptions. These results, in turn, enable a modular approach to security in the GGM and the AGM. As example applications, we use the UK assumption to prove knowledge soundness of Groth's zero-knowledge SNARK (EUROCRYPT 2016) and of KZG polynomial commitments (ASIACRYPT 2010) in the standard model, where for the former we reuse the existing proof in the AGM without hashing.

Jan von der Assen, Sandrin Raphael Hunkeler, Alberto Huertas Celdrán, Enrique Tomás Martínez Beltrán · 6 authors

No abstract is available for this record.

Mansi Gupta, Mohit Kumar, Renu Dhir

No abstract is available for this record.

Anas Harchi, Hicham Toumi, Mohamed Talea

As society increasingly relies on computers and automation, the challenge of developing secure applications, systems, and networks has become paramount. The complexity of modern networks and the proliferation of Internet of Things (IoT) devices have contributed to a surge in cyber threats facing individuals and organizations worldwide. Without effective collaboration, similar attacks can target multiple entities in rapid succession. While sharing cyber threat intelligence is often touted as a solution, privacy, trust, and traceability concerns persist. A novel distributed architecture is proposed to enhance IoT security to address these challenges. This solution relies on federated learning (FL) algorithms to establish a decentralized, autonomous system capable of detecting and characterizing attacks within a collaborative Cloud–SDN framework. Leveraging the strengths of Cloud computing and SDN, this architecture facilitates efficient and scalable data processing for IoT devices while safeguarding user privacy. By adopting FL, the model training process is decentralized, ensuring that sensitive data remains on the IoT devices, mitigating the risk of unauthorized access and data breaches.