Data Holdings Research Articles

10.B. Scientific session: The European Health Data Space (EHDS): A sea of change for health data in Europe

Abstract A wealth of data is generated every day throughout our health systems which, if leveraged well, can underpin the digital solutions and policies that will transform health and care. However, the complexity and divergence of rules, structures and processes within and across European Member States make it difficult to fully harness that potential at European scale. The European Health Data Space will address this challenge by empowering Europeans to access and control their health data and by offering a harmonised and simplified framework for the reuse of health data, including for research, innovation, and policymaking purposes. During the workshop, we will present an overview of the European Health Data Space framework and its implementation phases and discuss its benefits for targeted stakeholders, including citizens, researchers and policymakers in public health. Objectives 1. Emphasise the importance of leveraging health data for research, innovation and policymaking; 2. Present the European Health Data Space regulation and what actions are being undertaken to prepare for its implementation; 3. Discuss how access to their health data can enable patients and individuals play an active role in their own health journey; 4. Discuss what will be the main benefits for policymakers and researchers in public health resulting from enhanced access to health data. Introduction by Chair for presenting the European Health Data Space and the setting the scene for the session. Short presentations by speakers: 1. Implementing the EHDS: lessons from the EHDS2 pilot - Marianne Benderra, Health Data Hub, presenting his experience on developing and piloting the EHDS infrastructure for secondary use of health data, highlighting key project developments; 2. How can the EHDS enable participatory healthcare? - Henrique Martins, ISCTE, xShare project on how the xShare “Yellow Button” will empower individuals with access to and control over their health data and may foster the development of knowledge and tools necessary for prevention and management of their health; 3. What will the EHDS enable for research and policymaking? - Nienke Schutte, Scienciano, bringing the perspective of a leading European public health institute how the benefits the EHDS framework can bring for policymaking, research and innovation. 15 min Q&A session, taking questions from the audience, possibly also through Slido. Possible guiding questions: How can stakeholders contribute more broadly to the development of the EHDS? What support is needed to ensure that data users and data holders are prepared when the EHDS framework becomes operational? Conclusions 2 min by Chair. Speakers/Panelists Marianne Benderra French Health Data Hub, Paris, France Henrique Manuel Martins ISCTE, Lisbon, Portugal Nienke Schutte Sciensano, Brussels, Belgium

Summation-based Private Segmented Membership Test from Threshold-Fully Homomorphic Encryption

In many real-world scenarios, there are cases where a client wishes to check if a data element they hold is included in a set segmented across a large number of data holders. To protect user privacy, the client's query and the data holders' sets should remain encrypted throughout the whole process. Prior work on Private Set Intersection (PSI), Multi-Party PSI (MPSI), Private Membership Test (PMT), and Oblivious RAM (ORAM) falls short in this scenario in many ways. They either require data holders to possess the sets in plaintext, incur prohibitively high latency for aggregating results from a large number of data holders, leak the information about the party holding the intersection element, or induce a high false positive. This paper introduces the primitive of a Private Segmented Membership Test (PSMT). We give a basic construction of a protocol to solve PSMT using a threshold variant of approximate-arithmetic homomorphic encryption and show how to overcome existing challenges to construct a PSMT protocol without leaking information about the party holding the intersection element or false positives for a large number of data holders ensuring IND-CPA^D security. Our novel approach is superior to existing state-of-the-art approaches in scalability with regard to the number of supported data holders. This is enabled by a novel summation-based homomorphic membership check rather than a product-based one, as well as various novel ideas addressing technical challenges. Our PSMT protocol supports many more parties (up to 4096 in experiments) compared to prior related work that supports only around 100 parties efficiently. Our experimental evaluation shows that our method's aggregation of results from data holders can run in 92.5s for 1024 data holders and a set size of 2^25, and our method's overhead increases very slowly with the increasing number of senders. We also compare our PSMT protocol to other state-of-the-art PSI and MPSI protocols and discuss our improvements in usability with a better privacy model and a larger number of parties.

Australian medical regulations and strengthened continuing professional development (CPD): A policy implementation gap analysis with the specialist medical colleges

Introduction: Starting from 2023, Australian medical practitioners must meet new mandatory continuing professional development (CPD) requirements for registration renewal. In particular, each year they must complete 12.5 hours of educational activities and 25 hours of CPD activities related to reviewing performance and measuring outcomes, with a minimum of 5 hours each for reviewing performance and measuring outcomes activities, and they are encouraged to use patient health data and large eHealth datasets for better insights. Despite concerns regarding data accessibility and outcome measurement, specialist medical colleges are implementing these requirements. This study aims to explore factors influencing colleges’ efforts and provide recommendations for future research and action to enable data-driven CPD practices. Methods: A policy implementation gap analysis employing semi-structured interviews as a standalone method was conducted with participating colleges. Four colleges agreed to participate, with a total of 18 participants recruited from either the colleges’ staff who are responsible for interpreting and implementing CPD regulatory policies or the colleges’ CPD committees. Interviews were analysed using inductive and deductive thematic analysis. Data collection, data analysis and final results were reported using the COREQ checklist. Results: Three themes were identified from the interview analysis: (1) colleges' organisational structure and their roles and responsibilities within the Australasian CPD ecosystem, (2) challenges and enablers for policy implementation and (3) organisational needs and future actions at college level and necessary pre-conditions at ecosystemic level for strengthened CPD. Conclusions: Internal and external organisational factors are hindering colleges’ efforts to implement policy. More research, a reassessment of colleges’ roles, responsibilities and structure, and collaboration among colleges and data holders are necessary to enable full policy implementation and strengthened CPD practices.

Secure Framework Optimizes QAES Technique Used for Computing in the Cloud

Several consumers are concerned about the security of their data kept in the cloud, and many see local servers as a safer choice due to perceived control. However, cloud service companies frequently promise greater security procedures and employ dedicated security specialists, making data stored in the cloud potentially more secure. When the data holder expires, new files are saved in the cloud and encrypted with the QAES technology. The data owner then receives the encrypted data request. The sender and user both analyze data before sending encoded documents to the cloud. The key will be used to remove encryption from subsequent data. If users are unable to access the desired data, they should submit a fresh request. The most recent innovation focuses on combining an enhanced form of Quantum Key Distribution (QKD) with AES. This integration resulted in Quantum-AES (QAES), a novel quantum symmetric encryption scheme. QAES is based on the development of a quantum encryption key using dynamic quantum S-Boxes (DQS-Boxes), as opposed to the frequently utilized static S-Boxes. This strategy enhances security. Comparably, time is required to build files faster than they do now. This approach prevents brute force attacks since it uses the QAES algorithm, which provides additional security.

Transforming a Large-Scale Prostate Cancer Outcomes Dataset to the OMOP Common Data Model-Experiences from a Scientific Data Holder's Perspective.

To enhance international and joint research collaborations in prostate cancer research, data from different sources should use a common data model (CDM) that enables researchers to share their analysis scripts and merge results. The OMOP CDM maintained by OHDSI is such a data model developed for a federated data analysis with partners from different institutions that want to jointly investigate research questions using clinical care data. The German Cancer Society as the scientific lead of the Prostate Cancer Outcomes (PCO) study gathers data from prostate cancer care including routine oncological care data and survey data (incl. patient-reported outcomes) and uses a common data specification (called OncoBox Research Prostate) for this purpose. To further enhance research collaborations outside the PCO study, the purpose of this article is to describe the process of transferring the PCO study data to the internationally well-established OMOP CDM. This process was carried out together with an IT company that specialised in supporting research institutions to transfer their data to OMOP CDM. Of n = 49,692 prostate cancer cases with 318 data fields each, n = 392 had to be excluded during the OMOPing process, and n = 247 of the data fields could be mapped to OMOP CDM. The resulting PostgreSQL database with OMOPed PCO study data is now ready to use within larger research collaborations such as the EU-funded EHDEN and OPTIMA consortium.

Artificial intelligence based data curation: enabling a patient-centric European health data space.

The emerging European Health Data Space (EHDS) Regulation opens new prospects for large-scale sharing and re-use of health data. Yet, the proposed regulation suffers from two important limitations: it is designed to benefit the whole population with limited consideration for individuals, and the generation of secondary datasets from heterogeneous, unlinked patient data will remain burdensome. AIDAVA, a Horizon Europe project that started in September 2022, proposes to address both shortcomings by providing patients with an AI-based virtual assistant that maximises automation in the integration and transformation of their health data into an interoperable, longitudinal health record. This personal record can then be used to inform patient-related decisions at the point of care, whether this is the usual point of care or a possible cross-border point of care. The personal record can also be used to generate population datasets for research and policymaking. The proposed solution will enable a much-needed paradigm shift in health data management, implementing a 'curate once at patient level, use many times' approach, primarily for the benefit of patients and their care providers, but also for more efficient generation of high-quality secondary datasets. After 15 months, the project shows promising preliminary results in achieving automation in the integration and transformation of heterogeneous data of each individual patient, once the content of the data sources managed by the data holders has been formally described. Additionally, the conceptualization phase of the project identified a set of recommendations for the development of a patient-centric EHDS, significantly facilitating the generation of data for secondary use.

P3LS: Partial Least Squares under privacy preservation

Modern manufacturing value chains require intelligent orchestration of processes across company borders in order to maximize profits while fostering social and environmental sustainability. However, the implementation of integrated, systems-level approaches for data-informed decision-making along value chains is currently hampered by privacy concerns associated with cross-organizational data exchange and integration. We here propose Privacy-Preserving Partial Least Squares (P3LS) regression, a novel federated learning technique that enables cross-organizational data integration and process modeling with privacy guarantees. P3LS involves a singular value decomposition (SVD) based PLS algorithm and employs removable, random masks generated by a trusted authority in order to protect the privacy of the data contributed by each data holder. We demonstrate the capability of P3LS to vertically integrate process data along a hypothetical value chain consisting of three parties and to improve the prediction performance on several process-related key performance indicators. Furthermore, we show the numerical equivalence of P3LS and PLS model components on both a synthetic and a real-world dataset and provide a thorough privacy analysis of the former. Moreover, we propose privacy-preserving explained X- and Y-block variance computations for determining the contribution of each data holder to the federated process model as a basis to incentivize data federation and fair profit-sharing.

Generative artificial intelligence for distributed learning to enhance smart grid communication

Machine learning models are the backbone of smart grid optimization, but their effectiveness hinges on access to vast amounts of training data. However, smart grids face critical communication bottlenecks due to the ever-increasing volume of data from distributed sensors. This paper introduces a novel approach leveraging Generative Artificial Intelligence (GenAI), specifically a type of pre-trained Foundation Model (FM) architecture suitable for time series data due to its efficiency and privacy-preserving properties. These GenAI models are distributed to agents, or data holders, empowering them to fine-tune the foundation model with their local datasets. By fine-tuning the foundation model, the updated model can produce synthetic data that mirrors real-world grid conditions. The server aggregates fine-tuned model from all agents and then generates synthetic data which considers all data collected in the grid. This synthetic data can be used to train global machine learning models for specific tasks like anomaly detection and energy optimization. Then, the trained task models are distributed to agents in the grid to leverage them. The paper highlights the advantages of GenAI for smart grid communication, including reduced communication burden, enhanced privacy through anonymized data transmission, and improved efficiency and scalability. By enabling a distributed and intelligent communication architecture, GenAI introduces a novel way for a more secure, efficient, and sustainable energy future.

Linkage of routinely collected NHS data to evaluate liaison mental health services: challenges and lessons learned.

Liaison mental health services provide mental health care to patients in acute hospital settings. Evaluation of liaison services is challenging due to their heterogeneous organisation and delivery, high case throughput and varied patient case mix. We aimed to link routinely collected National Health Service data from secondary care settings, chosen for their service characteristics, to data from primary care to evaluate hospital-based liaison mental health services in England. We planned to compare patients referred to hospital-based liaison services with comparable patients in the same hospital not referred to liaison services and comparable patients in hospitals without any liaison services. We designed and enacted a methodology to link data from: (1) Hospital Episode Statistics, a database controlled by the National Health Service Digital and (2) ResearchOne, a primary care database controlled by The Phoenix Partnership. Obtaining approvals for the steps prespecified in the methodological protocol took 907 days. Enactment following approvals took 385 days. Data supplied from Hospital Episode Statistics contained 181,063 patients from 6 hospitals (mean=30,177, standard deviation=28,875.86) who matched the inclusion and exclusion criteria. Data supplied from ResearchOne contained 33,666 (18.6%) of these patients from the 6 hospitals (mean=5611, standard deviation=5206.59). Time required for approvals and enactment was attributable to slowness of data handling processes within each data holder and to resolution of technical and organisational queries between them. Variation in number of patients for which data was supplied between databases and between hospitals was attributable to coding inconsistencies and to the limited intersection of patient populations between databases and variation in recording practices between hospitals. Although it is technically feasible to link primary and secondary care data, the current system is challenging, complicated, unnecessarily bureaucratic, time consuming and costly. This limits the number of studies that could be conducted with these rich data sources. This article presents independent research funded by the National Institute for Health and Care Research (NIHR) Health and Social Care Delivery Research programme as award number 13/58/08.

Efficient privacy-preserving Gaussian process via secure multi-party computation

Gaussian processes (GPs), known for their flexibility as non-parametric models, have been widely used in practice involving sensitive data (e.g., healthcare, finance) from multiple sources. With the challenge of data isolation and the need for high-performance models, how to jointly develop privacy-preserving GP for multiple parties has emerged as a crucial topic. In this paper, we propose a new privacy-preserving GP algorithm, namely PP-GP, which employs secret sharing (SS) techniques. Specifically, we introduce a new SS-based exponentiation operation (PP-Exp) through confusion correction and an SS-based matrix inversion operation (PP-MI) based on Cholesky decomposition. However, the advantages of the GP come with a great computational burden and space cost. To further enhance the efficiency, we propose an efficient split learning framework for privacy-preserving GP, named Split-GP, which demonstrably improves performance on large-scale data. We leave the private data-related and SMPC-hostile computations (i.e., random features) on data holders, and delegate the rest of SMPC-friendly computations (i.e., low-rank approximation, model construction, and prediction) to semi-honest servers. The resulting algorithm significantly reduces computational and communication costs compared to PP-GP, making it well-suited for application to large-scale datasets. We provide a theoretical analysis in terms of the correctness and security of the proposed SS-based operations. Extensive experiments show that our methods can achieve competitive performance and efficiency under the premise of preserving privacy.

Survey on Stakeholder Cooperative Behavior for Designing Voluntary Medical Data Provision Motivation Mechanisms

The smooth provision of medical data is necessary to promote life science research utilizing artificial intelligence. However, data holders often do not voluntarily provide data. Previous literatures have stated several challenges of data sharing; however, studies on the perceptions of medical researchers are insufficient, especially in Japan. Aiming to explore data provision behavior in the actual Japanese context, we conducted a questionnaire survey on medical data use and data provision. The results clarified various obstacles to data sharing and revealed that whether the researcher is in an academic institution or a commercial company makes a difference.

A Dynamic Incentive Mechanism for Smart Grid Data Sharing Based on Evolutionary Game Theory

With the increasing popularization and application of the smart grid, the harm of the data silo issue in the smart grid is more and more prominent. Therefore, it is especially critical to promote data interoperability and sharing in the smart grid. Existing data-sharing schemes generally lack effective incentive mechanisms, and data holders are reluctant to share data due to privacy and security issues. Because of the above issues, a dynamic incentive mechanism for smart grid data sharing based on evolutionary game theory is proposed. Firstly, several basic assumptions about the evolutionary game model are given, and the evolutionary game payoff matrix is established. Then, we analyze the stabilization strategy of the evolutionary game based on the payoff matrix, and propose a dynamic incentive mechanism for smart grid data sharing based on evolutionary game theory according to the analysis results, aiming to encourage user participation in data sharing. We further write the above evolutionary game model into a smart contract that can be invoked by the two parties involved in data sharing. Finally, several factors affecting the sharing of data between two users are simulated, and the impact of different factors on the evolutionary stabilization strategy is discussed. The simulation results verify the positive or negative incentives of these parameters in the data-sharing game process, and several factors influencing the users’ data sharing are specifically analyzed. This dynamic incentive mechanism scheme for smart grid data sharing based on evolutionary game theory provides new insights into effective incentives for current smart grid data sharing.

Gesundheitsdaten: Von Nachbarländern lernen

Abstract The access and use of administrative health data in Germany is limited by several factors. Due to the large number and heterogeneity of data holders, data are usually only available in fragmented form, and access is often regulated in a non-transparent manner. Linkage of health data with other data sources is hardly possible for legal and logistic reasons. These limitations directly affect the ability of researchers in Germany to supply evidence-based policy advice and to conduct fundamental research. Against the backdrop of neighbouring countries, such as Denmark or Austria, giving access to and allowing the linkage of individual administrative data, we propose several measures to improve access to health data in Germany.

A federated and explainable approach for insider threat detection in IoT

An insider threat is a malicious action launched by authorized personnel inside the organization. Since insider actions may only leave a small digital footprint in the system, it is considered a major cybersecurity challenge in different applications. Along with the rapid growth of the Internet of Things (IoT) and the extensive attack surface in this technology, many concerns have been raised regarding the potential insider threats in IoT environments. Several studies have been conducted on Machine Learning (ML)-based insider threat detection solutions which are focused on the models’ performance while the trustability of these models is neglected. Trustworthy Learning refers to a new trend in ML that focuses on ways to ensure that the data collection and data analysis procedures in ML techniques follow ethical applications and are trustable to human users. This approach enforces the acceptance and successful adoption of ML-based solutions. This study aims to propose an improved trustworthy insider threat detection method that ensures two of the trustworthy learning requirements: Privacy and Explainability. The proposed solution protects the privacy of the utilized data and is capable of explaining why certain behaviors are detected as a threat. The proposed solution also leverages data collaboration between different data owners to increase the volume of data used in the training process and enhance the performance of the ML model. Experimental results show the proposed solution outperforms the learning models trained by individual data holders.

Privacy-Preserving Federated Recurrent Neural Networks

We present RHODE, a novel system that enables privacy-preserving training of and prediction on Recurrent Neural Networks (RNNs) in a cross-silo federated learning setting by relying on multiparty homomorphic encryption. RHODE preserves the confidentiality of the training data, the model, and the prediction data; and it mitigates federated learning attacks that target the gradients under a passive-adversary threat model. We propose a packing scheme, multi-dimensional packing, for a better utilization of Single Instruction, Multiple Data (SIMD) operations under encryption. With multi-dimensional packing, RHODE enables the efficient processing, in parallel, of a batch of samples. To avoid the exploding gradients problem, RHODE provides several clipping approximations for performing gradient clipping under encryption. We experimentally show that the model performance with RHODE remains similar to non-secure solutions both for homogeneous and heterogeneous data distributions among the data holders. Our experimental evaluation shows that RHODE scales linearly with the number of data holders and the number of timesteps, sub-linearly and sub-quadratically with the number of features and the number of hidden units of RNNs, respectively. To the best of our knowledge, RHODE is the first system that provides the building blocks for the training of RNNs and its variants, under encryption in a federated learning setting.

Collaborative training of deep neural networks for the lithium-ion battery aging prediction with federated learning

Accurate and reliable prediction of the future capacity degradation of lithium-ion batteries is crucial for their application in electric vehicles. Recent publications have highlighted the effectiveness of deep learning, in particular, in generating precise forecasts regarding the aging patterns. However, large quantities of training data covering various aging behaviors are required to train such models effectively. Collecting such a large database centrally is not feasible due to privacy and data communication restrictions of data owners, such as testing facilities or fleet operators. Federated learning provides a solution to this open issue. A framework, which incorporates federated learning into the training of a data-based battery aging model, is presented in this paper. The benefit of federated learning is that even data owners with sensible information can participate in a collaborative model training, since the model training is only conducted locally and all the data remains local and does not have to be disclosed. Thus, more data owners are likely to participate in this collaborative training. This will improve the prediction performance due to the enlarged dataset that can be utilized for the model training. This work shows that the prediction accuracy of the model trained with federated learning is only slightly worse than the prediction results obtained by the ideal case in which all aging data is stored in a central database. A sensitivity analysis is presented to prove the robustness of federated learning even if the datasets between participating data owners are highly imbalanced or exhibit different aging behaviors. Within exemplary scenarios, it is shown that individual data holders can reduce their prediction errors from MAPEmean=7.07% to MAPEmean=0.91% by participating in the proposed federated learning-based framework.

Novel Proposed Work for Empirical Word Searching in Cloud Environment

People's lives have become much more convenient as a result of the development of cloud storage. The third-party server has received a lot of data from many people and businesses for storage. Therefore, it is necessary to ensure that the user's data is protected from prying eyes. In the cloud environment, searchable encryption technology is used to protect user information when retrieving data. The versatility of the scheme is, however, constrained by the fact that the majority of them only offer single-keyword searches and do not permit file changes.A novel empirical multi-keyword search in the cloud environment technique is offered as a solution to these issues. Additionally, it prevents the involvement of a third party in the transaction between data holder and user and guarantees integrity. Our system achieves authenticity at the data storage stage by numbering the files, verifying that the user receives a complete ciphertext. Our technique outperforms previous analogous schemes in terms of security and performance and is resistant to inside keyword guessing attacks.The server cannot detect if the same set of keywords is being looked for by several queries because our system generates randomized search queries. Both the number of keywords in a search query and the number of keywords in an encrypted document can be hidden. Our searchable encryption method is effective and protected from the adaptive chosen keywords threat at the same time.

Turning data into research-ready data

ObjectivesGovernments acquire extensive data holdings and face increasing pressure to make these available as record-level microdata for research. However, turning data into research-ready data (RRD) is not a straightforward exercise. We demonstrate how even in simple cases researcher involvement can bring substantial rewards for effective RRD development.
 MethodsThis paper reports on an ADRUK-funded project to take a dataset originally collected by the Office for National Statistics for official statistics (the UK Annual Survey of Hours and Earnings, ASHE), formally review its microanalytical characteristics, link it to Census 2011 data, and prepare a new ‘research ready dataset’ with appropriate documentation and coding. This should have been straightforward as the datasets had already been widely used as research microdata. However, the involvement of academic researchers in the production of research-ready data led to many important new insights.
 ResultsThe research programme had 3 aims: testing assumptions about the data; reviewing data quality; and adding value.
 Because of its sampling model, ASHE is assumed to have random non-response both longitudinally and in cross section. The research team showed that was untrue: there was higher attrition than expected, and both longitudinal and cross-sectional non-response appeared non-random..
 The data quality review showed further concerns about the accuracy of some geographical indicators, and some variables of opaque provenance; in contrast, we confirmed the accuracy of administrative variables created by ONS.
 As well as being important for researchers, these findings have the potential for significant effects on official statistics produced from the source data, enhancing the value of the source data.
 Finally, value was added from new variables which reflected the team’s wide research interests
 ConclusionOften in government the assumption is that creating RRDs is a matter of creatign files and giving access to the researchers. Insights from our work show that the deep involvement of the research community can bring rewards for both data holders and researchers. For RRDs, researcher-led construction is vital.

Implementing the CARE Principles for Indigenous Data Governance in Biodiversity Data Management

Indigenous data governance is a critical aspect of upholding Indigenous rights and fostering equitable partnerships in biodiversity research and data management. An estimated 80% of the planet’s biodiversity exists on Indigenous lands (Sobrevila 2008), and the majority of Indigenous data derived from specimens taken from Indigenous lands are held by non-Indigenous entities and institutions. The CARE Principles (Collective benefit, Authority to control, Responsibility, and Ethics) are designed to guide the inclusion of Indigenous peoples in data governance, and increase their access to and benefit from data (Carroll et al. 2020). This talk will share emerging tools and resources that can be leveraged to implement the CARE Principles within repositories and institutions that hold Indigenous data. This talk highlights two primary tools to promote Indigenous data governance in repositories: a phased framework to guide third-party holders of Indigenous data through foundational learning and concrete steps to apply the CARE principles in their respective settings, and the CARE criteria, an assessment tool by which researchers and institutions can evaluate the maturity of CARE implementation, identify areas for improvement, and allow other entities such as funders and publishers to evaluate CARE compliance. a phased framework to guide third-party holders of Indigenous data through foundational learning and concrete steps to apply the CARE principles in their respective settings, and the CARE criteria, an assessment tool by which researchers and institutions can evaluate the maturity of CARE implementation, identify areas for improvement, and allow other entities such as funders and publishers to evaluate CARE compliance.

An Improved Federated Learning-Assisted Data Aggregation Scheme for Smart Grids

In the context of rapid advancements in artificial intelligence (AI) technology, new technologies, such as federated learning and edge computing, have been widely applied in the power Internet of Things (PIoT). When compared to the traditional centralized training approach, conventional federated learning (FL) significantly enhances privacy protection. Nonetheless, the approach poses privacy concerns, such as inferring other users’ training data through the global model or user-transferred parameters. In light of these challenges, this research paper introduces a novel privacy-preserving data aggregation scheme for the smart grid, bolstered by an improved FL technique. The secure multi-party computation (SMC) and differential privacy (DP) are skillfully combined with FL to combat inference attacks during both the learning process and output inference stages, thus furnishing robust privacy assurances. Through this approach, a trusted third party can securely acquire model parameters from power data holders and securely update the global model in an aggregated way. Moreover, the proposed secure aggregation scheme, as demonstrated through security analysis, achieves secure and reliable data aggregation in the electric PIoT environment. Finally, the experimental analysis shows that the proposed scheme effectively performs federated learning tasks, achieving good model accuracy and shorter execution times.