Nur Shahidah Mohammad, Md. Mafizur Rahman, Md. Abdur Razzaque
Federated Learning (FL) enables multiple entities to collaboratively train models without sharing sensitive data, but it faces critical privacy, security, and efficiency challenges in healthcare intrusion detection systems. These issues are intensified by adversarial attacks, non-IID data, and the need for real-time performance. Existing FL methods struggle with gradient inversion, model poisoning, Sybil attacks, and high computational overhead, limiting their effectiveness in secure and scalable healthcare applications. This work proposes the PrivacyPreserving Personalized Federated Learning Intrusion Detection in Healthcare applications (P3FL-HIDS), integrating Byzantinerobust aggregation, gradient masking, and Zero-Knowledge Proof based authentication. Key features include strong adversarial resilience, protection of privacy against gradient inversion, personalized model adaptation for heterogeneous data, and secure participant authentication. Additional contributions include a dual-network training approach, adaptive clustering for personalization, and optimized secure communication for real-time healthcare scenarios. Experimental results on a Brain Tumor magnetic resonance imaging (MRI) dataset show that P3FLHIDS outperforms state of the art works in terms of accuracy, resilience, and resistance.
Efficient extraction and integration of pharmacokinetic (PK) data from scientific literature is critical for informed decision-making in drug development. Prior knowledge of PK parameters, particularly from similar compounds, supports first-in-human dosing, parameter estimation, and compound screening, ultimately helping to reduce attrition in clinical trials. While recent natural language processing (NLP) efforts have focused on extracting PK data from unstructured text, these approaches often overlook more comprehensive PK information and essential contextual metadata, which are usually reported in tables. Despite the prevalence and value of these tables, no previous work has systematically addressed the automated extraction of PK data from them. This thesis presents a novel NLP pipeline for identifying, extracting, and structuring PK data from scientific tables. The work addresses a key gap by targeting tables as a rich and underutilised source of PK information. The thesis is structured around four main components. First, a classification system combining supervised learning and prompt-based approaches is developed to retrieve PK-relevant tables from full-text biomedical articles. Second, heuristic and neural named entity recognition approaches are designed to extract PK parameters and associated metadata from table cells, including dose, species, study population, route of administration, units, and other contextual qualifiers. Third, an entity linking pipeline, including rule-based and zero-shot methods, is developed to normalise extracted data to a standardised PK ontology. Finally, the full pipeline is utilised to construct a large-scale PK database from PubMed Open Access articles. The database is evaluated through systematic sampling and manual quality assessment, and proof-of-concept analyses demonstrate how the extracted data can be used to characterise literature-wide reporting trends and explore comparative pharmacological questions. The results of this thesis demonstrate that automated PK table mining is both feasible and scalable, significantly accelerating the curation of high-quality datasets for pharmacometrics modelling. This work presents new open-source annotated corpora, domain-specific NLP methodologies, and practical tools for structuring PK literature, thereby opening the door to scalable, data-driven approaches in early drug development.
Shrutika Singh, Anton Alyakin, Daniel Alexander Alber, Jaden Stryker · 12 authors
The performance of Large Language Models (LLMs) on multiple-choice question (MCQ) benchmarks is frequently cited as proof of their medical capabilities. We hypothesized that LLM performance on medical MCQs may in part be illusory and driven by factors beyond medical content knowledge and reasoning capabilities. To assess this, we created a novel benchmark of free-response questions with paired MCQs (FreeMedQA). Using this benchmark, we evaluated three state-of-the-art LLMs (GPT-4o, GPT-3.5, and LLama-3-70B-instruct) and found an average absolute deterioration of 39.43% in performance on free-response questions relative to multiple-choice (p = 1.3 * 10 -5 ) which was greater than the human performance decline of 22.29%. To isolate the role of the MCQ format on performance, we performed a masking study, iteratively masking out parts of the question stem. At 100% masking, the average LLM multiple-choice performance was 6.70% greater than random chance (p = 0.002) with one LLM (GPT-4o) obtaining an accuracy of 37.34%. Notably, for all LLMs the free-response performance was near zero. Our results highlight the shortcomings in medical MCQ benchmarks for overestimating the capabilities of LLMs in medicine, and, broadly, the potential for improving both human and machine assessments using LLM-evaluated free-response questions.
Open access
Artificial Intelligence in Healthcare and Education
Early diagnosis of cardiac abnormalities depends on accurate classification of heart sounds, but centralized training methods run the danger of violating patient privacy. We thus propose a privacy-preserving and reliable heart sound abnormality detection system combining Blockchain Technology with Federated Learning (FL). Training is spread among seven clients, each simulating an independent data source, using a preprocessed dataset from the PhysioNet Challenge 2016 to enable distributed learning without sharing raw data. CNN-LSTM model using FedAvg achieved the best performance: 94\% accuracy, 0.90 precision, 0.96 recall, and an AUC of 0.98 among five deep learning architectures evaluated with FedAvg and FedProx strategies. Along with metadata including client ID and round number, SHA-256 hashes of local and global model weights were recorded on a local Ethereum blockchain following every communication round to guarantee model integrity. The hash of the final model is revalidated against the blockchain to confirm authenticity prior to deployment. It then guarantees safe, distributed, clinically valuable AI-based diagnostics by real-time classification of heart sounds as normal or abnormal.
Federated learning (FL) offers a distributed approach for the collaborative training of machine learning models across decentralized clients while safeguarding data privacy. This characteristic makes FL well suited for privacy-sensitive fields such as healthcare and finance. However, addressing the heterogeneity caused by nonindependent and identically distributed (non-IID) data remains a significant challenge for traditional FL methods. To address these issues, the enhancing clustered federated learning with adaptive similarity (AS-CFL) algorithm, which dynamically forms client clusters based on model update similarity and uses a forward-incentive mechanism to improve collaborative training efficiency among similar clients, is proposed in this study. Experimental results on the MNIST and EMNIST datasets reveal that compared with baseline methods such as the CFL, IFCA, and FedAvg models, the AS-CFL algorithm achieves faster convergence—reducing the number of communication rounds by approximately 20%—while maintaining competitive accuracy, demonstrating its effectiveness in heterogeneous FL scenarios.
Traditional health insurance models depend on fixed policy structures that often lack personalization and transparency. These static systems are slow to respond to individual health behaviors, which leads to inefficient premium calculations and rigid coverage terms. In this paper, we propose a decentralized and smart framework that allows real-time negotiation of health insurance coverage using dynamic non-fungible tokens (dNFTs) and AI-powered risk profiling. Our system uses Ethereum smart contracts to represent health insurance policies as dNFTs that change based on individual health metrics collected from wearable devices or electronic health records. A builtin machine learning engine evaluates user risk profiles in realtime, updating the dNFT metadata as needed. Insurers interact with the system through a decentralized marketplace, where they bid to provide personalized coverage terms based on the live health profile encoded in the token. The negotiation process is trustless, transparent, and automatic using smart contracts, which removes intermediaries and cuts down processing time. We show a working prototype deployed on the Ethereum testnet and assess it using simulated user data. The results indicate a noticeable improvement in policy adaptability, personalization, and claim settlement time. This approach has the potential to transform health insurance by making it dynamic, data-driven, and fully decentralized.
Machine Learning in Healthcare
Artificial Intelligence in Healthcare and Education
Blockchain and Federated Learning (FL) provide a strong framework for distributed, privacypreserving machine learning in the medical field. In order to provide safe and effective model training, this framework assists in handling sensitive patient data from lung disease diagnosis, such as CT scans, X-rays, and clinical records. The proposed approach improves distributed machine learning security, privacy, and integrity, particularly in delicate fields like healthcare. Contributions from other datasets help the model get better, but patient data is kept private and blockchain guarantees the integrity of the updates to the model. ZeroKnowledge Proofs (ZKP) guarantee that customers can demonstrate the accuracy of their model upgrades without disclosing any personal information. FLBC- ZKP uses cryptographic proofs to remove this requirement for confidence. FLBC-ZKP models exhibit competitive accuracy rates in healthcare applications, guaranteeing confidentiality and privacy without compromising predictive performance. Contri- butions from other datasets improve the model, but patient information is kept confidential and the blockchain ensures the accuracy of model updates. Compared to regular FL, FLBC-ZKP delivers superior privacy and security through blockchain and ZKP, making it particularly suitable for sensitive healthcare data, while maintaining high accuracy. The accuracy data throughout federated learning rounds for a different approach, FLBC-ZKP slightly surpasses the other methods as the number of rounds increases.
In this report, we compare the performance of our UltraGroth-based zero-knowledge machine learning framework Bionetta to other tools of similar purpose such as EZKL, Lagrange's deep-prove, or zkml. The results show a significant boost in the proving time for custom-crafted neural networks: they can be proven even on mobile devices, enabling numerous client-side proving applications. While our scheme increases the cost of one-time preprocessing steps, such as circuit compilation and generating trusted setup, our approach is, to the best of our knowledge, the only one that is deployable on the native EVM smart contracts without overwhelming proof size and verification overheads.
This paper presents a novel health data analysis platform for improved individualized risk forecasting of permanent pacemaker implantation (PPI) following transcatheter aortic valve replacement (TAVR) procedures. Specifically, we introduce XGBoost-federated adaptive interpolation transfer learning (XG-FedAIT)—a platform that integrates heterogeneous hospital data sets via adaptive output-level interpolation and performance-weighted model ensembling. This approach facilitates federated learning across institutions with different feature spaces and prediction targets, such as time-to-event and binary models, eliminating structural and semantic mismatches prevalent in regular federated transfer learning. To ensure secure, privacy-preserving, and general data protection regulation (GDPR)-compliant data exchange, we propose a dual-chain blockchain architecture integrating proof of authority (PoA), zero-knowledge proofs (ZKPs), and chain-specific smart contracts, and off-chain encrypted storage through Filecoin. Experimental evaluation proves that the PrimaryChain handles over 300 tx/s with a median latency of 500 ms, and the SecondaryChain offers 99.8% data availability and decentralized access control. The system is scalable to handle up to 2,500 transactions/hour, and the federated learning pipeline classifies PPI risk with an F1-score of 0.85 and AUROC of 0.91. These results support the effectiveness of our system in delivering real-time, regulation-compliant, and clinically actionable cardiac care in distributed environments.
Open access
Artificial Intelligence in Healthcare and Education
Blockchain is an emerging technology that is being used to create innovative solutions in many areas, including healthcare. Nowadays healthcare systems face challenges, especially with security, trust, and remote data access. As patient records are digitized and medical systems become more interconnected, the risk of sensitive data being exposed to cyber threats has grown. In this evolving time for healthcare, it is important to find a balance between the advantages of new technology and the protection of patient information. The combination of blockchain–InterPlanetary File System technology and conventional electronic health record (EHR) management has the potential to transform the healthcare industry by enhancing data security, interoperability, and transparency. However, a major issue that still exists in traditional healthcare systems is the continuous problem of remote data unavailability. This research examines practical methods for safely accessing patient data from any location at any time, with a special focus on IPFS servers and blockchain technology in addition to group signature encryption. Essential processes like maintaining the confidentiality of medical records and safe data transmission could be made easier by these technologies. Our proposed framework enables secure, remote access to patient data while preserving accessibility, integrity, and confidentiality using Ethereum blockchain, IPFS, and group signature encryption, demonstrating hospital-scale scalability and efficiency. Experiments show predictable throughput reduction with file size (200 → 90 tps), controlled latency growth (90 → 200 ms), and moderate gas increase (85k → 98k), confirming scalability and efficiency under varying healthcare workloads. Unlike prior blockchain–IPFS–encryption frameworks, our system demonstrates hospital-scale feasibility through the practical integration of group signatures, hierarchical key management, and off-chain erasure compliance. This design enables scalable anonymous authentication, immediate blocking of compromised credentials, and efficient key rotation without costly re-encryption.
The increase in demand of data driven decision making in sensitive fields like healthcare and finance requires machine learning frameworks that maintain strict data privacy and follow regulations. Federated Learning (FL) provides a decentralized way to train models. It allows multiple organizations to learn together from distributed datasets without sharing raw data. But, traditional FL methods, such as Federated Averaging (FedAvg), face issues in real world situations. These issues arise from different data distributions among clients and the risk of information leaks from shared model updates. In this research study, we introduce a new federated learning framework with two main innovations: First the adaptive aggregation strategy that adjusts client contributions based on how stable they are and their quality, and second an optional differential privacy module at the server to make sure privacy guarantees. We tested the framework on two publicly available datasets: a heart disease dataset from the University of California, Irvine (UCI) repository and a large financial dataset from Kaggle. This simulates collaboration between hospitals and financial institutions. Experimental results show that our adaptive aggregation method boosts model accuracy by up to 4.2% compared to FedAvg, while still performing well even with differential privacy applied. The model achieves an AUC of 0.93 and an F1 score of 0.891, with minimal communication overhead. These results confirm the framework’s strength and its ability to support the ethical use of Artificial Intelligence in regulated and data sensitive areas. They also recommend it can scale effectively across larger federated networks.
Privacy-Preserving Technologies in Data
Artificial Intelligence in Healthcare and Education
The secure management of Electronic Health Records (EHRs) in a cloud environment poses many challenges, and guaranteeing the scalability of a secure solution to manage the huge amounts of data and its privacy remains an open problem. Although the existing encryption methods offer strong security, the widespread adoption of asymmetric-key protocols is limited because of the lack of computational efficiency compared with practical applications. That is, the efficiency of computational time required for an encryption or decryption step, and the privacy-preserving verification of the computation result, are not balanced due to the volume of the data. In order to overcome the drawbacks of existing encryption methods, we propose a novel approach to integrating Zero-Knowledge Proofs (ZKPs) with a Multi-Layer Merkle Tree (MLMT) to achieve a scalable, privacy-preserving, guarantee-of-integrity and publicly-verifiable solution for managing the EHR while ensuring patients’ privacy. It proposes the utilization of MLMT to build a hierarchical data verification structure for massive data, significantly improving the computational efficiency. It also employs ZKP to enable verifier to verify the validity of data without revealing any record information, which is vital for the data management in the healthcare sector. The authors compared the proposed model with existing approaches which adopt AES-256 encryption and typical Merkle Tree-based solutions and demonstrate its superior scalability and privacy-preserving ability while ensuring controllable computational overhead. The results demonstrate that ML MMT ZKP provides the best balance between privacy, integrity and scalability reaching lower overheads and shorter verification times than other traditional approaches. This work constitutes a step forward in the development of cryptographic solutions for EHRs and provides a framework for real-time verifiable information in the healthcare domain.
Alejandro Guerra-Manzanares, Omar El-Herraoui, Michail Maniatakos, Farah E. Shamout
One of the key challenges of collaborative machine learning, without data sharing, is multimodal data heterogeneity in real-world settings. While Federated Learning (FL) enables model training across multiple clients, existing frameworks, such as horizontal and vertical FL, are only effective in ‘ideal’ settings that meet specific assumptions. Hence, they struggle to address scenarios where neither all modalities nor all samples are represented across the participating clients. To address this gap, we propose BlendFL, a novel FL framework that seamlessly blends the principles of horizontal and vertical FL in a synchronized and non-restrictive fashion despite the asymmetry across clients. Specifically, any client within BlendFL can benefit from either of the approaches, or both simultaneously, according to its available dataset. In addition, BlendFL features a decentralized inference mechanism, empowering clients to run collaboratively trained local models using available local data, thereby reducing latency and reliance on central servers for inference. We also introduce BlendAvg, an adaptive global model aggregation strategy that prioritizes collaborative model updates based on each client’s performance. We trained and evaluated BlendFL and other state-of-the-art baselines on three classification tasks using a large-scale real-world multimodal medical dataset and a popular multimodal benchmark. Our results highlight BlendFL’s superior performance for both multimodal and unimodal classification. Ablation studies demonstrate BlendFL’s faster convergence compared to traditional approaches, accelerating collaborative learning. Overall, in our study we highlight the potential of BlendFL for handling multimodal data heterogeneity for collaborative learning in real-world settings where data privacy is crucial, such as in healthcare and finance.
Weiming Xie, Zhaomin Yao, Xiaozhou Bai, Lang Mai · 10 authors
Nowadays, the global digital transformation of healthcare is advancing rapidly with the help of technologies such as electronic medical records, telemedicine, and mobile medical applications. However, there are still challenges in EMR interoperability, security, and data exchange. To address these existing limitations, This study proposes a voice electronic medical record system driven by artificial intelligence and blockchain, which is designed to improve clinical records and nursing coordination. This system adopts a dedicated deep learning architecture. It transcribe the conversations between doctors and patients into text, and then uses natural language processing to extract the relevant medical information. At the same time, it also provides diagnostic prompts, which can reduce the risk of misdiagnosis. Doctors can view and edit these summaries generated by artificial intelligence. Then safely record them on the decentralized crypto blockchain ledger. With federated learning, the model can be continuously improved in multiple centers without infringing on data privacy. This solution integrates automatic speech recognition, distributed ledger technology, and collaborative deep learning, aiming to enhance the EMR efficiency, security, data integrity, and care continuity of medical institutions. The combination of blockchain technology and artificial intelligence technology holds great potential. It can transform fragmented health data into portable and interoperable records under patient control, thus bringing strategic advantages to the health system that is undergoing a comprehensive digital transformation.
The integration of machine learning (ML) in healthcare has unlocked transformative potential in disease prediction, personalized treatment, medical imaging, remote patient monitoring, and genomic data analysis. However, the sensitive nature of medical data introduces critical concerns regarding patient privacy, data security, and regulatory compliance. This chapter presents a comprehensive overview of privacy-preserving machine learning approaches tailored for healthcare applications, with a focus on technical frameworks, real-time implementations, and regulatory alignment. It explores the use of advanced techniques such as federated learning, differential privacy, homomorphic encryption, and zero-knowledge proofs to safeguard patient information while maintaining model utility. The chapter also addresses domain-specific challenges in processing real-time health data streams and implementing privacy-aware algorithms in resource-constrained environments. By bridging the gap between technical innovation and clinical applicability, this work emphasizes the importance of secure, scalable, and ethically aligned ML solutions in modern healthcare ecosystems. The discussion was contextualized within current legal frameworks and highlights future directions for research and implementation to ensure trust, transparency, and resilience in data-driven medical systems.
The complexities of healthcare data, including privacy concerns, imbalanced datasets, and interoperability issues, necessitate innovative machine learning solutions. Swarm Learning (SL), a decentralized alternative to Federated Learning, offers privacy-preserving distributed training, but its reliance on blockchain technology hinders accessibility and scalability. This paper introduces a \textit{Simplified Peer-to-Peer Swarm Learning (P2P-SL) Framework} tailored for resource-constrained environments. By eliminating blockchain dependencies and adopting lightweight peer-to-peer communication, the proposed framework ensures robust model synchronization while maintaining data privacy. Applied to cancer histopathology, the framework integrates optimized pre-trained models, such as TorchXRayVision, enhanced with DenseNet decoders, to improve diagnostic accuracy. Extensive experiments demonstrate the framework's efficacy in handling imbalanced and biased datasets, achieving comparable performance to centralized models while preserving privacy. This study paves the way for democratizing advanced machine learning in healthcare, offering a scalable, accessible, and efficient solution for privacy-sensitive diagnostic applications.
Sharmila Agnal, Venkata Ramana K., Dinesh Kumar S, Ashwin Ulagappan G
This Project presents a safe and decentralized framework for handling electronic medical records (EMRs) using blockchain technology. The suggested framework guarantees data integrity, confidentiality, and availability through the use of blockchain's distributed ledger system. It allows healthcare providers, patients, and approved staff to safely access and update medical records in real-time while ensuring rigorous privacy controls. This solution overcomes the problems with current EMR systems, including data fragmentation, security risks, and interoperability problems. By leveraging smart contracts and blockchain's cryptographic security, this framework can transform healthcare data management, improve patient care, and enable efficient collaboration among medical professionals.
Blockchain and machine learning are transformative technologies with the potential to significantly enhance the healthcare sector by addressing critical challenges such as operational efficiency, data security, and privacy. As a distributed ledger technology, blockchain enables secure and decentralized data sharing among healthcare systems, facilitating the safe exchange of sensitive patient information between hospitals, diagnostic labs, pharmaceutical companies, and patients. Blockchain applications in healthcare can improve patient privacy protection, electronic health record (EHR) management, and counterfeit drug tracking within pharmaceutical supply chains. When integrated with machine learning, blockchain fosters the development of advanced, patient-centric healthcare solutions. Its decentralized architecture ensures data integrity, prevents manipulation, and enhances security, while AI-powered analytics enable personalized medicine, real-time decision-making, and operational cost reduction. This synergy is expected to revolutionize healthcare by enabling applications such as secure medical data storage, pharmaceutical supply chain transparency, remote patient monitoring, and fraud prevention in clinical trials. By leveraging the complementary strengths of these technologies, healthcare systems worldwide can improve patient outcomes, enhance efficiency, and establish a more transparent and resilient infrastructure.
Blockchain technology is transforming industries like finance, supply chain, governance, and healthcare. This paper analyzes blockchain architecture, applications, challenges, and performance. Key attributes—decentralization, immutability, transparency, and security—enable secure peer-to-peer transactions without intermediaries. Consensus mechanisms such as Proof of Work (PoW) and Proof of Stake (PoS) are examined for their trade-offs in scalability, energy efficiency, and security. In healthcare, blockchain addresses security and interoperability issues in centralized Personal Health Record (PHR) systems. Solutions using Ethereum, Hyperledger, smart contracts, and IPFS enhance Electronic Health Record (EHR) management by improving data integrity, privacy, and access control while reducing costs. A containerized microservices architecture further enhances scalability. Blockchain performance, evaluated using the BLOCKBENCH framework, highlights gaps in transaction throughput compared to traditional databases. Despite scalability, interoperability, and regulatory challenges, ongoing research focuses on optimizing consensus mechanisms, integrating database principles, and improving healthcare interoperability, advancing blockchain’s real-world applications
Patrick Silva, Patrick J. Silva, Patrick A. Silva, Patrick A. Silva · 5 authors
Cloud and distributed computing, code repositories, and large language models are democratizing the less computationally intensive use cases of artificial intelligence (AI) in medicine. The convergence and democratization of these powerful tools promises to mobilize and utilize humanity’s knowledge and data, at least the knowledge bases and data that are readily available in the public commons. Healthcare represents a challenge due to fragmentation of the data fabric and governance mechanisms intrinsic to that sector of the economy. Privacy laws, stewardship practices, and the fragmented nature of the patient data journey (medical record silos) create cumbersome impediments to health data sharing, particularly longitudinal patient‐level data. Consequently, obtaining the data necessary to train and operationalize AI in many healthcare and clinical genomics use cases limits the promise of these new technologies in addressing complexities in healthcare. We posit that trust, provenance, and fitness of health data and transaction costs represent challenges that blockchain ledgers and smart digital contracts might address. Here, we present frameworks from some of the great economic thinkers that might help address some of the stewardship and agency issues inherent to health data sharing. Our goal is to promote a more equitable and patient‐centric healthcare data fabric to address current challenges of healthcare.
Open access
Artificial Intelligence in Healthcare
Machine Learning in Healthcare
Artificial Intelligence in Healthcare and Education
Recently, with the prevalent adoption of blockchain in the financial system, there has been an increasing of anomaly activities such as ponzi schemes, gambling and phishing fraud on Ethereum platforms, and an effective account classification method is urgently required. The existing account classification methods on Ethereum with high accuracy require a learning system to be trained with balanced datasets. However, the distribution of annotated labels for account identities published on third-party sites is relatively imbalanced. Therefore, in this paper, We propose a EthGAN framework which includes a high-dimensional node feature representation module and a few-shot account data augment module to improve the accuracy and robustness at imbalanced datasets. The high-dimensional node feature representation module captures features from statistical, temporal, and transaction structure, and the few-shot account data augmentation module based on generative adversarial network models generate few-shot samples to improve the diversity and representativeness of the training datasets. We conduct extensive experiments to evaluate the performance of our proposed EthGAN framework on real-world Ethereum transaction data. The average classification effect of our method is 10+% higher than that of existing methods. Experimental results demonstrate that our method outperforms state-of-the-art methods in Ethereum account classification.