Data Sharing Research Articles

Large language models (LLMs) are often trained on undisclosed data. This practice has intensified debates about transparency, copyright compliance, and reproducibility. From this lens, this paper systematically reviews methodologies to detect shared training data across LLMs. More specifically, our review spans five methodological families: 1) lexical/semantic overlap metrics, which compare output similarity but struggle with knowledge convergence; 2) memorization analysis, which identifies verbatim regurgitation of rare training examples yet risks extracting copyrighted content; 3) temporal alignment by leveraging models’ knowledge cutoffs to infer shared data timelines; 4) adversarial susceptibility correlation to measure shared failure modes under attack; and 5) synthetic fingerprinting by embedding detectable artifacts in training data. Our review reveals several critical gaps: existing methods are highly siloed, with little cross-community dialogue, evaluation inconsistency epidemics, and ethical risks are understudied and potentially violate data privacy laws. We propose a taxonomy of auditing techniques, an approach to conduct the same, and conduct a case study to showcase its merit. Finally, this review highlights an under-explored knowledge gap and sets a roadmap for future research directions.

Abstract Brazil hosts the highest fish diversity globally, underscoring the critical role of Brazilian Scientific Fish Collections (BSFCs) in biodiversity research and conservation. Despite their importance, BSFCs face persistent challenges. This study introduces the CARDUME network, the first nationwide initiative to assess BSFCs and propose strategies to enhance their role in ichthyology. Data from 74 collections across 25 Brazilian federal units reveal that most are affiliated with public universities and house 8 502 992 catalogued specimens in 906 890 lots, including over 2600 primary types. Although digitization efforts are underway in 92% of BSFCs, only 37% currently share data online. Genetic resources are severely limited, with tissue samples available for just 3% of specimens. Significant funding disparities exist, with 70% of collections receiving less than USD170 annually. Gender inequality is also notable, with more male-curated than female-curated collections. Survey responses highlighted issues such as staff shortages, inadequate facilities, and limited resources for fieldwork and collection curation. CARDUME aims to address these gaps through ­collaboration, professional training, improvement of data quality, and increase of shared data, while also advocating for recognition by Brazilian institutions, funding agencies, and the government. O resumo em português está disponível no material suplementar.

Background/Objectives: Precision medicine is transforming care for children with cancer, but raises new challenges. We explored parents’, oncologists’ and scientists’ perspectives on three aspects of a precision medicine trial for poor prognosis childhood cancer: data sharing, requests for additional tumor biopsies, and confidentiality. Methods: Data were collected through PRISM-Impact, a psychosocial sub-study within the Zero Childhood Cancer Program’s PRISM trial. Parents completed questionnaires at enrolment and one year later, and an optional interview after receiving their child’s trial results. Bereaved parents completed a questionnaire six months after bereavement (T1B). Oncologists and scientists were interviewed one year following trial commencement. Quantitative data were analyzed descriptively, and qualitative data thematically. Results: Parents (n = 126) considered additional tumor biopsies acceptable when risks were low and their child or oncologist supported the request. Oncologists (n = 26) emphasized weighing risk–benefit, ensuring parents felt fully informed, and research value. Most parents supported data sharing (≥89–96%), including after bereavement, despite potential privacy concerns. Parents supported overseas and interstate testing, and scientists having access to identifiable health information. Scientists (n = 10) found working with identifiable data emotionally challenging. Conclusions: Parents, oncologists, and scientists showed high acceptance of procedural aspects of precision medicine. Future trials should address privacy concerns and ensure informed consent recognizes that parents’ high acceptability of procedures may be linked to their hopes for benefit, reinforcing the need for informed consent.

Patients with hepatocellular carcinoma (HCC) and Medicaid insurance on the liver transplantation (LT) waitlist face higher risks of waitlist dropout, though mechanisms behind this disparity remain unclear. We aimed to assess differences in the use of locoregional therapy (LRT) based on insurance status and whether these differences contribute to waitlist disparities. We conducted a retrospective cohort study using Organ Procurement and Transplantation Network/United Network for Organ Sharing (OPTN/UNOS) data on adult patients (≥ 18years) waitlisted with standardized HCC Model for End-stage Liver Disease (MELD) exceptions from 1/1/2015 to 12/31/2022. Mixed effects multiple variable logistic regression models were used to evaluate the association between insurance status and LRT receipt, adjusting for key clinical and HCC-related variables. Patients with Medicaid had higher odds (OR: 1.09; 95% CI: 1.01, 1.18) of receiving LRT compared to patients with private insurance. Additionally, when comparing waitlist time following HCC MELD exception approval for our cohort, Medicaid patients experienced longer median waitlist time (206days; IQR 88-371) compared to privately insured patients (182days; IQR 69-313) (p < 0.001). Contrary to expectations, Medicaid patients were more likely to receive LRT than those with private insurance. These findings highlight the importance of further investigating contributing factors that facilitate these outcomes.

The size of molecular dynamics (MD) trajectories remains a major obstacle for data sharing, long-term storage, and ensemble analysis at scale. Existing solutions often rely on frame subsampling or reduced atom representations, which limit the utility of shared data sets. Here, we present MDZip, a neural compression framework based on convolutional autoencoders trained per system to reconstruct atomic trajectories with high geometric fidelity from compact latent representations. MDZip achieves over 95% reduction in storage size across a diverse benchmark of proteins, protein-peptide complexes, and nucleic acids. Despite operating in a physics-agnostic manner, the reconstructed trajectories accurately preserve ensemble-level features, including RMSD fluctuations, pairwise distance distributions, radius of gyration, and projections onto principal and time-lagged independent components. A residual (skip-connected) autoencoder variant consistently improves reconstruction accuracy and reduces outliers. While local structural deviations can impair energetic fidelity, short energy minimization partially recovers physically reasonable conformations. This framework enables customizable compression-accuracy trade-offs and supports a modular workflow for sharing latent representations, decoder models, and reconstruction protocols. MDZip offers a scalable solution to current storage limitations, facilitating broader dissemination of MD data without sacrificing essential dynamical information.

In modern high-performance computing (HPC) and large-scale data processing environments, the efficient utilization and scalability of memory resources are critical determinants of overall system performance. Architectures such as non-uniform memory access (NUMA) and tiered memory systems frequently suffer performance degradation due to remote accesses stemming from shared data among multiple tasks. This paper proposes LACX, a shared data migration technique leveraging Compute Express Link (CXL), to address these challenges. LACX preserves the migration cycle of automatic NUMA balancing (AutoNUMA) while identifying shared data characteristics and migrating such data to CXL memory instead of DRAM, thereby maximizing DRAM locality. The proposed method utilizes existing kernel structures and data to efficiently identify and manage shared data without incurring additional overhead, and it effectively avoids conflicts with AutoNUMA policies. Evaluation results demonstrate that, although remote accesses to shared data can degrade performance in low-tier memory scenarios, LACX significantly improves overall memory bandwidth utilization and system performance in high-tier memory and memory-intensive workload environments by distributing DRAM bandwidth. This work presents a practical, lightweight approach to shared data management in tiered memory environments and highlights new directions for next-generation memory management policies.

Securing analytics on shared data is important but expensive. Analyzing datasets from multiple data owners can yield valuable insights [1, 2, 3, 4, 5] but poses significant security risks. Even within enterprises - our primary focus - precautions are necessary when handling data across subsidiaries and geographic regions [6, 7]. Existing security solutions based on Trusted Execution Environments (TEEs) [8, 9], fully homomorphic encryption [10], and structured encryption [11] offer strong protections, albeit in a physically centralized manner. For more decentralization, there are exciting approaches based on Secure Multi-Party Computation (MPC) [12] that do not need a trusted third party nor merging datasets at a central location. Recent projects [6, 13, 14, 15] show that MPC can reduce the risk of leaks for analytics on shared data under stronger security guarantees. However, MPC queries are often impractically slow, requiring orders of magnitude more computation and communication than plain-text or TEEbased query execution.

In the era of digital transformation in the banking sector, there has been an increasing practice of sharing customer data with third parties, such as service providers or vendors. This practice poses legal challenges, particularly concerning the fulfillment of the principles of necessity and proportionality in the protection of personal data. This study aims to analyze the implementation of these two principles in the collaborative practices between banks and vendors regarding the protection of customers’ personal data. This normative juridical research employs a conceptual and statutory approach, using legal materials obtained from national and international regulations, academic journals, and best practices in the banking sector. The findings indicate the need for clear and comprehensive internal bank policies on personal data protection in third-party data processing, serving as a guideline to ensure compliance with personal data protection principles.

The Gulf of Mexico’s (also recognized by the United States government as the Gulf of America; herein referred to as “the Gulf”) valuable and diverse marine, coastal, and estuarine environments sustain many habitats, species, and economically important fisheries that are vulnerable to open ocean and coastal acidification (OOCA), including shellfish, coral reefs, and other carbonate reefs and seafloor. OOCA poses an economic threat to the Gulf’s economy, which is estimated to have a combined value of $2.04 trillion (US) per year across Cuba, Mexico and the United States (U.S.). Scientists from Cuba, Mexico, and the U.S. co-organized and co-hosted the first Gulf International Ocean Acidification Summit on Oct. 18-19, 2022 in Mérida, Yucatan, Mexico to exchange information and begin development of a new tri-national network to address the socioeconomic and ecological impacts of OOCA in the Gulf based on common needs. The meeting included representatives from government agencies, universities, research institutes, non-governmental organizations, and was sponsored by the Furgason Fellowship of the Harte Research Institute at Texas A&amp;amp;M University-Corpus Christi. Discussions focused on each country’s challenges, including known and potential socioeconomic vulnerabilities and biological and ecosystem responses to OOCA. Shared priorities were identified for observational, biological, environmental needs, socioeconomic research, outreach, and communications. Priority geographic locations for the study and short and long-term monitoring of OOCA were identified based on the group’s knowledge of oceanographic conditions and vulnerable regions. Longer-term actions that will help support multinational collaborations include: identifying shared data and information platforms; standardizing chemical and biological sampling methodologies; coordinating communications with regulatory agencies and resource managers; and coordinating monitoring activities, collaborative research projects, and tri-national comparisons and synthesis of findings. We present guidance from this effort for an integrated, multinational approach to understanding the causes and consequences of OOCA in the Gulf.

To date, no shared guidelines have been approved for the diagnosis and management of low bone mineral density (BMD), especially in early infancy. Therefore, there is an increasing demand for new methodologies to allow the assessment of bone health status in this specific cohort, which is exposed to several risk factors (e.g. maternal vitamin D deficiency, pregnancy-associated diseases, preterm birth and comorbidities, low birth weight, intrauterine growth restriction). Currently, the assessment of BMD in newborn and infants relies mainly on serum and urinary biochemical markers, in association with several technologies to measure bone mineral content, such as dual-energy X-ray absorptiometry (DXA) and quantitative ultrasound (QUS) being traditionally used, despite many limitations. More recently, Radiofrequency Echographic Multi-Spectrometry (REMS) emerged as a promising tool in clinical practice for screening and monitoring BMD. Due to the radiation-free technology, an extremely ease of use, low costs, an excellent degree of sensitivity, specificity, and reproducibility, REMS technology has proven to be the gold standard technique in sensitive populations such as pregnant women, newborns and infants, allowing mass extended screening strategies. However, to date no validate cut-off reference for REMS in paediatric age are available. Future longitudinal studies on REMS methodology are needed to build reference standards and new shared algorithms, combining biochemical and instrumental data, for the diagnosis, management and treatment of decreased BMD before and after birth.

The rapid development of e-commerce has led to a significant increase in the risk of sensitive data leakage, such as user transaction records, payment details, and identity information. Information security issues are becoming increasingly severe, especially in scenarios that rely on centralized cloud storage architectures, facing core challenges such as single point of failure leakage risks, lack of fine-grained access control in cross organizational data sharing, high encryption computing costs, rigid existing access control policies, and the contradiction between privacy protection and data sharing efficiency. In view of this, research has proposed and validated an innovative solution that integrates blockchain technology. Specifically, the study proposes a data security protocol based on attribute proxy re encryption and a data security sharing model based on blockchain. The data security protocol adopts ciphertext policy attribute encryption and introduces attribute proxy re encryption technology to re encrypt ciphertext through proxy nodes, solving the privacy protection problem of data transmission and sharing. The data security sharing model utilizes blockchain to store shared data, designs detailed access control policies and data encryption and decryption control mechanisms, and constructs a data security sharing system under a multi-layer architecture. In the experimental section, a simulated attack environment was constructed to test data security protocols. The results showed that compared with protocols such as SRAAP, TEE Oracle, Pedersen link Schnorr, etc., the blockchain data sharing protocol designed in this study achieved a ‘high’ level in terms of resistance to 51% attacks and resistance to general attacks (better than the ‘medium’ level of SRAAP protocol). The encryption time for privacy identity data was only 2.84ms (lower than SRAAP’s 5.20ms and TEE Oracle’s 7.46ms), and the maximum encryption time was controlled at 32ms. When processing a large amount of private data, the indexing time of the encryption control algorithm did not exceed 0.4 seconds, far lower than B-SEM algorithm’s 0.79s and CAB algorithm’s 0.91s. Moreover, the average memory occupation of the research scheme was only 18.54%, significantly lower than F-SEM algorithm. In the verification of blockchain data sharing platforms, compared with data sharing schemes such as BDAE, BF, RAISE, etc., the data control error rate of the research model does not exceed 3%, the accuracy of privacy data transmission is close to 95%, the data search and encryption performance is excellent (average search time of 1.35ms, shortest encryption time), and the access control cost is lower than other models. The research method combines blockchain technology and attribute proxy re encryption technology to effectively improve the privacy protection and control performance of shared data. It is significantly better than existing mainstream solutions in terms of computing efficiency, resource consumption, security strength, and control accuracy, effectively solving the inherent defects of centralized cloud storage, providing enhanced privacy protection, access control, and secure sharing capabilities for e-commerce data, ensuring data integrity and confidentiality, and reducing the threat risk of privacy sensitive data.

BackgroundLittle is known regarding patient attitudes toward automatic sharing of race, ethnicity, and language (REL) data in health care settings despite the universal practice of data sharing across health care institutions and providers.ObjectiveThis study aims to assess public comfort with disclosing and automatically sharing REL data in health care settings and understand the social factors associated with these attitudes.MethodsUsing the 2022 DataHaven Community Wellbeing Survey from 1196 adult Connecticut residents, we examined factors associated with public comfort with disclosing and automatically sharing REL data across health care settings. We generated unadjusted and adjusted logistic models to examine associations between factors and responses to the data-sharing questions.ResultsMost residents surveyed were White (n=873, 73%), followed by African American or Black (n=167, 14%), Asian or Native Hawaiian or other Pacific Islander (n=31, 2.6%), multiracial (n=31, 2.6%), and American Indian or Alaska Native (n=12, 1%). The majority of respondents were not Hispanic or Latino (n=1051, 87.9%). More than half of respondents reported excellent or very good self-rated health (SRH; n=635, 53.1%), and most participants reported almost always trusting their health care provider (n=939, 78.5%). Most participants reported being willing to share race and ethnicity data at a hospital or clinic (n=1008, 84.3%) and REL data automatically (n=947, 79.2%) in health care settings. Hispanic or Latino (adjusted odds ratio [AOR] 0.049, 95% CI 0.25-0.94) and multiracial (AOR 0.32, 95% CI 0.14-0.76) respondents were less likely to be willing to disclose race and ethnicity data compared to those who were not Hispanic or Latino and who were White, respectively. Individuals who sometimes trust health care providers (AOR 0.57, 95% CI 0.35-0.94) or rarely/never (AOR 0.35, 95% CI 0.15-0.85) were less likely to be willing to disclose race and ethnicity data than those who almost always trust health care providers. African American or Black (AOR 0.46, 95% CI 0.29-0.72) and American Indian or Alaska Native (AOR 0.18, 95% CI 0.04-0.75) individuals were less likely to be willing to share REL data automatically than White individuals. Those who sometimes trust health care providers (AOR 0.48, 95% CI 0.31-0.74) or rarely/never trust health care providers (AOR 0.25, 95% CI 0.11-0.56) were less likely to be willing to share REL data automatically than those who almost always trust health care providers. Those with poor/fair SRH versus very good/excellent SRH were less likely to be willing to share REL data automatically (AOR 0.54, 95% CI 0.34-0.85).ConclusionsRacial and ethnic identity, SRH, and trust in health care providers affect willingness to share REL information with providers and other health systems.

With population aging, more adults across Europe face competing demands of working for pay and caring for elderly family members. Associated tradeoffs are expected to be negative, gendered, and vary across contexts with different levels of gender equality, public support for eldercare, and work-family balance. Using SHARE data from 2004–20, this study investigates how unpaid caregiving to independently living parents relates to labor supply among mature working-age (50–64) men and women across Europe. Results find limited tradeoffs between unpaid caregiving and labor supply, even where public support for eldercare is low. Caregiving associates with men’s and women’s employment and full-time work in similar ways. Gender differences nevertheless exist in both paid work and caregiving across Europe, especially in Continental and Southern Europe. These differences are established before midlife and build up across the life course and should be addressed when designing policies for longer working lives in Europe.

Abstract. This paper presents D-TECH, an open-source platform designed to support the documentation, visualization, and collaborative management of cultural heritage assets through a distributed, modular, and semantically driven environment. Developed under the D-TECH project and funded by the DTC Lazio initiative, the platform incorporates state-of-the-art technologies for 3D acquisition, Augmented Reality (AR), Virtual Reality (VR), and Geographic Information Systems (GIS), and ensures semantic interoperability aligned with ARCO standards. D-TECH addresses critical institutional challenges in the management and reuse of 3D heritage data, providing tools for storage, visualization, and sharing within a modular and federated environment. This platform enables public institutions to manage their digital assets in-house, ensuring full control, visualization, annotation, and dissemination through a federated network of nodes. The paper details the software architecture, data workflows, 3D visualization tools, and use scenarios, highlighting its role in facilitating collaborative heritage work and unlocking the potential of existing 3D data archives.

The Challenges of Access and Sharing of Linguistic Data from the VARSUL Base Sample

Rationale and ObjectivesLoneliness has emerged as a global public health concern with broad implications for mental and physical health across populations. Yet existing tools and frameworks tend to individualize what is often a structurally embedded and socially patterned experience. This scientific session aims to: 1. Reframe loneliness as both individual and collective phenomenon across the lifespan; 2. Introduce and validate new metrics that move beyond standard loneliness scales; 3. Share national-level trend data and forecasts on loneliness in Finland to 2040; 4. Explore adolescent loneliness in relation to mental health and substance use, highlighting social exclusion and protective factors in family and school contexts.Added ValueThis session synthesizes three complementary lines of research-scale development, population-level modeling, and adolescent mental health-to build a multi-dimensional understanding of loneliness. It will bring theoretical clarity and methodological innovation to a field dominated by reductive metrics. Key contributions include: • A novel scale capturing collective loneliness at multiple social levels; • Empirical evidence of sharply rising loneliness across age groups in Finland, with long-term projections; • Identification of differential patterns of adolescent loneliness and the moderating roles of institutional support. These insights are timely for policymakers and researchers responding to WHO's recommendations on strengthening social connection as a pillar of public health.Coherence Between PresentationsEach presentation addresses loneliness at a different scale and life stage, offering layered and non-redundant insights: • Beattie et al. introduce and validate the Collective Loneliness Scale, theorizing this neglected type of loneliness as illbeing arising from alienation from social groups (groups, community, society); • Parikka et al. provide a population-level analysis showing past and future loneliness trends by gender and age-groups up to 2040; • Tunkkari et al. focus on adolescents, differentiating between social, emotional, and ostracism-based loneliness, and examining protective buffers in familial and institutional contexts. Together, the presentations trace loneliness as a developmental and structural issue, offering converging evidence and distinct lenses.Workshop FormatThis session follows the Scientific session format. We propose: • Three presentations of ∼12 minutes each; • A short integrative commentary (1 speaker, 6 minutes) to draw out thematic connections; • A moderated Q&A and audience interaction (∼15 minutes), emphasizing cross-cutting policy and research implications; • All presenters have agreed to structure their contributions to maximize thematic clarity and audience engagement.Key messages• Loneliness is increasing across age groups; projections to 2040 highlight urgent need for action.• New measures and tailored interventions are needed to address both interpersonal and collective loneliness.

ABSTRACTSecuring the trustworthiness, privacy, and legitimacy of shared medical data in Wireless Body Area Network (WBAN) is a primary concern. Hence, a blockchain technology‐based secure medical data storage scheme is developed in this paper. This developed model includes four primary phases. Before initializing, the WBAN data are collected. In the first phase, the user authentication is verified. For this purpose, the user's iris images are aggregated. These iris images are subjected to the Residual Attention Network (RAN). From the RAN, the user is authorized, and then security keys are given to the authorized user. Only after verifying the authentication of the user, the healthcare data is allowed to be stored in the blockchain. In the second phase, data sanitization takes place. The obtained WBAN medical data are sanitized using a data sanitization process with the optimal keys obtained from the Fusion of Golden Eagle and Eurasian Oystercatcher Optimization Algorithm (FGE‐EOOA). Here, the data are encrypted by employing the Rivest‐Shamir‐Adleman (RSA) approach, and then encrypted medical data are stored in the blockchain. This ensures multi‐step data security, which allows secure storage of WBAN healthcare data in the blockchain. While retrieving the stored data, the user authentication is verified on the user side, as well as in the same RAN model. This is the third phase of the developed model. When the user is proven to be an authorized one, the stored data in the blockchain corresponding to that particular user is retrieved. Using the data restoration process, which is the fourth phase of the developed model, the actual medical data is retrieved. If the user is unauthorized, then no access is provided to them. This ensures a multi‐level of security for storing and retrieving data from the blockchain. The security offered by this model is evaluated and validated by contrasting and comparing it with other conventional data transfer methods.

This study aimed to provide updated nationwide data on outcomes in heart transplantation for amyloid cardiomyopathy (ACM) compared with other cardiomyopathy etiologies. United Network for Organ Sharing data for patients over 18 who underwent heart transplant for ACM, and other cardiomyopathies including DCM, ICM, RCM, and CHD were reviewed from 2001 through 2022. The data were analyzed with a propensity score-matched analysis comparing ACM patients with transplant for the cardiomyopathies to measure primary outcomes, including 5-year, 10-year, and all-cause mortality accounting for differences at baseline. Of 21 457 heart transplant recipients who met eligibility criteria, there was a 100% increase in the number of heart transplants for ACM from 2010 to 2020 (p < 0.001). The unadjusted mortality did not differ significantly among ACM, DCM, ICM, RCM, and CHD groups, and 5-year mortality was comparable between ACM and DCM patients. ACM patients had a significantly lower incidence of postoperative stroke (p = 0.044), and risk factors for ACM transplant mortality were identified as mechanical ventilation at the time of transplant (HR: 3.8, p = 0.023) and older donor age (HR: 1.1, p = 0.015). The number of heart transplants for ACM has increased in recent years, and overall outcomes in these ACM transplants have been similar compared to those for other cardiomyopathies. Despite historical concerns regarding poor prognosis and amyloid recurrence, carefully screened ACM patients can benefit from heart transplantation. Further research and optimization of ACM patient eligibility criteria alongside investigation of concurrent adjunctive therapies could optimize treatment of ACM with heart transplantation.

Risks and protective factors for cognitive maintenance in men and women: A secondary analysis of the longitudinal SHARE data.

To solve the problems of data silos and information asymmetry in traditional supply chain information security sharing, this article combines Convolutional Neural Networks (CNN) and blockchain consensus algorithms, analyzes data and uses blockchain for secure sharing, so that all parties can obtain and verify data in real time, improve the overall operational efficiency of the supply chain, and promote information transparency and sharing efficiency. CNN can be used to analyze data in the supply chain. Training on real digital images ensures data privacy and improves the accuracy and efficiency of data processing. Blockchain technology can be introduced into supply chain information sharing to ensure the immutability and transparency of data. This article introduces a federated learning (FL) mechanism to improve consensus algorithms, which improves the efficiency of model training. Among them, each link in the FL process is rigorously verified and recorded through the consensus mechanism of blockchain, ensuring the security and reliability of the entire process. This article adopts an improved consensus algorithm, PoDaS (Proof of Data Sharing), whose core idea is to use the computational consumption generated during FL as proof of workload. The specific steps include: local model training and uploading, model update verification shield, and model update aggregation. The PoDaS algorithm combines the advantages of PoW (Proof of Work) and PoS (Proof of Stack) to ensure the fairness of the consensus mechanism and reduce the waste of computing resources. By comparing and analyzing the block time and model accuracy of three algorithms, the superiority of PoDaS algorithm in block time and model accuracy was verified. The experimental results show that the PoDaS algorithm is significantly better than the PoW algorithm in terms of block generation time, and slightly better than the PoS algorithm. In terms of model accuracy, the PoDaS algorithm is significantly superior to traditional PoW and PoS algorithms. Its model accuracy reaches 96.00%, reflecting the effectiveness and practicality of the PoDaS consensus algorithm in the sharing of supply chain information security.