The work, led by Zijian Yao of Hangzhou Normal University together with collaborators at Westlake University, including Qiqi Liu and corresponding author Xiumei Li, addresses a tension that has stymied privacy-preserving federated learning for years. Personalized federated learning, or PFL, allows each client to adapt a model to its own local data distribution—for example, a hospital whose patients differ demographically from those of every other hospital. Typical PFL methods split the model into two parts: a subset of parameters that is sent to the server for global aggregation, and the rest that stays local for personalization. The trouble is that the choice of what to share is usually made heuristically, and the shared portion can still leak private information. Moreover, if one tries to hide the shared parameters using homomorphic encryption—a cryptographic technique that allows computation directly on encrypted values—the communication and computation costs can balloon to the point of being unusable, particularly when different clients choose to share different parts of the model.

FedFIHE attacks both problems simultaneously, and the first half of its solution draws on a classical concept from information theory and statistical learning: Fisher information. The Fisher information matrix measures how sensitive a model’s likelihood is to changes in each of its parameters. In practical terms, a parameter with high Fisher information is one the model relies on heavily to fit the data; disturbing it would sharply change performance. Conversely, parameters with low Fisher information contribute less to the model’s fit on any particular client’s data. The researchers exploit this by having each client compute Fisher-information scores across its network and use them to adaptively partition the model. Parameters deemed most informative in a global sense become candidates for sharing and aggregation, while the remainder are retained locally for personalization. Because the partitioning is guided by the data’s actual statistical structure rather than arbitrary rules, the shared subset tends to carry the generalizable knowledge that benefits all clients while the private, idiosyncratic details stay behind.

This Fisher-guided partitioning does more than improve accuracy under heterogeneous data—it also reduces the surface area for privacy attacks. Since the parameters most strongly tied to any individual client’s private data distribution are preferentially kept local, an inquisitive server sees a filtered version of each model that carries less identifying signal. The team reports that this design effectively mitigates privacy leakage compared with approaches in which clients share large or arbitrarily chosen portions of their models, while still delivering personalized performance that is better than or comparable to strong baseline methods.

The second half of the framework confronts the encryption bottleneck directly. Homomorphic encryption schemes, such as the lattice-based methods commonly used in privacy-preserving machine learning, allow a server to add ciphertexts together—which is exactly what federated averaging requires—without ever decrypting them. The catch is that encrypted values are bulky and expensive to process. If clients could agree on a single shared set of parameters, encrypting just that subset would be manageable. But in personalized federated learning, each client’s Fisher-guided selection may differ, and the union of many clients’ selections can approach the whole model. Naively, the server would have to encrypt and aggregate nearly everything, destroying the efficiency gains the partitioning was supposed to provide.

FedFIHE’s answer is a packet-level selective encrypted aggregation scheme with a clever twist of voting. Each client first packs its model parameters into fixed-size packets—containers that bundle many individual weights together. The client then applies a binary mask indicating which packets contain parameters selected for sharing, ranks the packets, and uploads only the top-κ packets by its selection score, along with those binary masks. When the packets arrive, the server performs what the authors call mask voting: it tallies, across all clients, which packet positions were most frequently selected. Only those consensus positions—the packets that a sufficient number of clients chose—are aggregated in ciphertext and broadcast back. The rest are simply never touched by the encrypted pipeline.

The elegance of this design lies in how it handles disagreement. In previous selective-sharing approaches, inconsistent sparse selections across clients meant that the union of selected positions grew with the number of participants, and the server’s encrypted workload grew with it. Mask voting caps this cost by construction: no matter how differently the clients choose, only the most popular packet positions enter the encrypted aggregation, so the server’s workload stays bounded and predictable. The binary masks themselves are lightweight to transmit, and packing many parameters per packet amortizes the overhead of ciphertext expansion over hundreds or thousands of individual weights. The result, according to the authors, is encrypted training that remains efficient even under high sparsity—situations where clients share only a small fraction of their model—precisely the regime where naive schemes become prohibitively expensive.

The research team validated FedFIHE through extensive experiments on both natural image benchmarks and medical datasets, a deliberate choice reflecting the framework’s target use cases. Healthcare is arguably the most compelling arena for federated learning: hospitals cannot pool patient records, yet diagnostic models would benefit enormously from learning across institutions. The medical data experiments showed that FedFIHE preserves its privacy and efficiency advantages on realistic clinical distributions, where data heterogeneity between clients is typically severe. Across the evaluation suite, the method mitigated privacy leakage while achieving personalized accuracy that matched or exceeded strong baselines, and it retained its robustness even when the degree of selective sharing was pushed to high sparsity levels.

The significance of the work extends beyond any single benchmark. Federated learning has been deployed at scale by major technology companies for mobile keyboard prediction and other on-device applications, and national health initiatives have explored cross-institutional model training for rare-disease detection. In all of these settings, the honest-but-curious threat model is not hypothetical: gradient-inversion attacks have demonstrated that raw training data can sometimes be reconstructed from shared updates with startling fidelity. Privacy researchers have responded with a toolbox of defenses—differential privacy adds calibrated noise, secure multi-party computation distributes trust, and homomorphic encryption hides values entirely—but each carries costs in accuracy, coordination complexity, or computational overhead. FedFIHE’s contribution is to show that a statistically principled sharing policy can shrink the encrypted workload enough to make homomorphic encryption practical for personalized models, rather than treating privacy and utility as a simple trade-off.

The technical machinery is worth appreciating in its own right. Fisher information has historically served as the foundation for natural gradient optimization and for understanding which weights matter most when pruning neural networks. Repurposing it as a privacy-aware routing mechanism—deciding which parameters travel to the server and which stay home—is a novel application of a decades-old tool. Meanwhile, the mask-voting protocol echoes ideas from consensus systems: instead of forcing clients to agree in advance on what to share, the system tolerates disagreement and resolves it after the fact, cheaply, in the clear. Only the agreed-upon ciphertext positions incur the heavy cost of homomorphic computation. The combination means the strongest cryptographic protection is applied where the most clients have converged, and the sparser, more personal corners of each model remain local by design.

The authors are candid that the framework is aimed at privacy-sensitive applications such as healthcare, and the experimental evidence on medical datasets supports that positioning. Because FedFIHE is compatible with heterogeneous data—each client personalizes its own model after receiving the aggregated global portion—it does not require the unrealistic assumption that all participants draw from the same distribution. That flexibility, combined with bounded communication costs, makes the approach a candidate for real deployments where hospitals, research consortia, or mobile networks must cooperate without exposing patient or user data to the coordinating server.

The study was supported in part by the China-Croatia bilateral science and technology cooperation project, the Hangzhou Joint Fund of the Zhejiang Provincial Natural Science Foundation of China, and the Xinmiao Talent Program of Zhejiang Province. Published open access, the research is available to any institution weighing whether privacy-preserving collaborative learning is ready to leave the laboratory. For a field that has spent a decade watching the gap between cryptographic promises and practical constraints, FedFIHE offers evidence that the gap can be narrowed—parameter by parameter, packet by packet, and vote by vote.

The new framework, called MAQDRL — short for Multi-Agent Federated Deep Q-Learning with QuadTree spatial partitioning — is described in an open-access paper published in the journal Cognitive Computation. Led by Aditya Kumar Raj, Anurag Sharma, and K. Jairam Naik, the team set out to solve two bottlenecks that have long plagued autonomous vehicle routing in complex urban settings: the latency and privacy problems of streaming raw sensor data to centralized servers, and the computational explosion of collision checking as fleets grow. Their answer combines three ideas that had rarely been fused together before — lightweight local reinforcement learning, federated averaging of model parameters, and a data structure borrowed from computational geometry.

At the heart of MAQDRL is a deceptively simple division of labor. Each agent, whether a simulated vehicle or a future delivery robot, trains its own compact Deep Q-Network using only local observations: its grid position, the elapsed timestep, its Euclidean distance to its goal, and the average distance to nearby agents within a defined collision boundary. No raw trajectories, sensor streams, or replay-buffer samples ever leave the agent. Instead, after every five episodes, each agent sends only its neural network parameters — a mere 2,629 trainable weights, roughly 10.27 kilobytes in 32-bit floating point — to a mobile edge computing aggregator, where they are blended through federated averaging into a shared global policy that is then redistributed to the fleet. In a ten-agent scenario, an entire synchronization round costs just over 0.20 megabytes of communication, and ten rounds of training consume approximately 2 megabytes in total — a figure the authors describe as modest enough to fit comfortably within the constraints of roadside mobile edge computing units.

The second pillar is the QuadTree, a classic spatial indexing structure that recursively subdivides a two-dimensional region into four quadrants whenever the number of agents in a cell exceeds a preset capacity. In MAQDRL’s simulated world — a 100 × 100 grid populated by ten agents with randomly assigned start positions and unique goals — the tree is rebuilt at every timestep to reflect the agents’ latest locations. A lazy merge policy collapses sibling nodes whose combined agent count falls below half the capacity threshold, preventing the structure from shattering into excessive fragments in sparse regions. The payoff is dramatic: querying which agents lie within a given collision boundary drops from quadratic cost to logarithmic time, O(log N), allowing each agent to check only its spatially relevant neighborhood rather than the entire population. The team’s complexity analysis puts the overall per-timestep cost at O(N(d + log N)), where d is the size of the local policy network — a far cry from the O(N²) dependence that plagues baseline methods such as MADDPG, MAPPO, and MARC, which rely on centralized critics, joint action spaces, or global replay buffers.

The training recipe itself is anchored in a compact neural architecture the authors call SAF-QNet. A five-dimensional state vector — x and y coordinates, goal distance, timestep, and local proximity — feeds into a fully connected hidden layer of 64 ReLU neurons, followed by a 32-neuron hidden layer and a five-output head producing Q-values for the discrete action set: up, down, left, right, or stay. Careful preprocessing underpins stability: state features are min-max normalized to the range [0, 1] so that no single variable dominates gradient calculations, actions are one-hot encoded so the network never mistakes “up” for being numerically adjacent to “down,” and rewards are standardized by subtracting the mean and dividing by the standard deviation of the reward vector. Gaussian noise augmentation with a variance of 0.01 is applied to state trajectories to harden policies against perturbation. The reward function blends a positive signal for reaching the goal, a penalty for collisions, and a proximity term that decays with distance to the target, nudging agents toward short, energy-efficient paths.

The evaluation environment was deliberately designed to be demanding. The benchmark is derived from CVRPTW-style routing instances — the Capacitated Vehicle Routing Problem with Time Windows — although the authors are careful to note that their setup addresses decentralized point-to-point spatial navigation rather than the full constraint-laden route feasibility problem. Because each agent starts from a different position, pursues a different goal, encounters different neighbors, and accumulates different collision and reward histories, the local training data are naturally non-IID, the heterogeneous setting that has historically been the Achilles’ heel of federated learning. Yet the federated averaging procedure appears to tame this heterogeneity: by aggregating parameters every five episodes, the framework reduces the variance of individual updates and preserves the contraction behavior of the Bellman operator, yielding learning curves as smooth as those of a single-agent DQN trained on stable data.

The headline numbers are striking. MAQDRL achieved a task success rate of 97.8 percent, comfortably ahead of MAPPO at 94.8 percent, MADDPG at 95.2 percent, and MARC at 92.2 percent. Collision frequency fell by 73.08 percent relative to MAPPO and 66.13 percent relative to MADDPG, while energy consumption — measured as the summed Euclidean distance traveled by all agents — dropped by 56.52 percent and 55.98 percent against the same two baselines. Over 100 training episodes, MAQDRL recorded just 21 total collisions, compared with 62 for MADDPG, 78 for MAPPO, 98 for MARC, and 192 for a random-policy control. Cumulative rewards told the same story: MAQDRL finished at approximately 125, against 85 for MADDPG, 68 for MAPPO, 60 for MARC, and barely 20 for the random policy. Average energy consumption per episode settled around 265.98 kilojoules for MAQDRL, well below the roughly 604 to 612 kilojoules consumed by MADDPG and MAPPO and the erratic 714.93 kilojoules of the random baseline.

Ablation experiments confirmed that neither of the two innovations alone accounts for the gains. A DQN paired with QuadTree indexing but no federated synchronization showed fast early rewards but plateaued prematurely, unable to propagate successful strategies across agents. A Federated DQN without the QuadTree produced smooth, steady improvement but lacked the spatial awareness needed to respond quickly to proximity conflicts. Only the full integration — local DQN learning, periodic federated averaging, QuadTree-based spatial indexing, and reward shaping — delivered the best trade-off across stability, safety, and energy efficiency, maintaining energy levels roughly 20 to 30 percent below its ablated counterparts. Hyperparameter sweeps reinforced the picture: a conservative learning rate of 0.001 yielded the highest total reward of 4,736 with only 21 collisions, while faster rates of 0.1 and 0.01 triggered aggressive overcorrection, collision spikes of up to four per episode, and energy peaks above 300 kilojoules. Similarly, a slow exploration decay of 0.995 — which preserves exploration longer before committing to exploitation — consistently produced better rewards, fewer collisions, and lower energy use than decay rates of 0.8 or 0.9.

The implications extend well beyond simulation. The authors argue that the framework’s small model footprint, low synchronization overhead, and linear aggregation complexity make it a natural fit for MEC-assisted edge deployments, where roadside units must aggregate updates from many vehicles within tight bandwidth and compute budgets. Keeping raw sensor streams local also sidesteps regional data-privacy regulations and the sub-100-millisecond latency thresholds that centralized LiDAR and V2X transmission can violate. The theoretical analysis offers an intuitive explanation for why the combination works: by confining each agent’s interactions to a small, logarithmically sized spatial neighborhood, the QuadTree dampens the non-stationarity that other agents introduce into the environment, producing smoother temporal-difference targets and a tighter effective contraction factor for the Bellman operator — the mathematical engine that underpins stable reinforcement learning.

The team is candid about the limits. The current experiments ran in a simplified obstacle-free grid with ten agents, and the communication and aggregation analysis is analytical rather than a hardware-level benchmark of inference latency, memory footprint, or energy draw on constrained edge devices. Extreme local density, adversarial agents, noisy localization, and obstacle-rich environments may all demand additional mechanisms, from robust aggregation to asynchronous synchronization and personalized federated learning. Future work will extend the framework toward full CVRPTW-aware routing, test it on realistic traffic simulators such as SUMO and CityFlow, scale the agent population beyond fifty, and deploy the trained policies on single-board computers to validate real-world feasibility. Still, as urban fleets of autonomous vehicles edge closer to reality, the message of this study is clear: the fastest way for machines to avoid crashing into each other may be to learn separately, share sparingly, and always know exactly where to look.

The core of this pioneering research lies in the development of a federated deep learning architecture—a decentralized machine learning paradigm where multiple clients or devices collaboratively train a model without sharing raw data. This foundation is crucial in addressing privacy concerns that have long plagued centralized data aggregation systems. By enabling local data processing and only sharing model updates, the framework safeguards sensitive information while still harnessing the collective knowledge of a vast distributed network.

To enhance the efficacy of fake news detection, the researchers ingeniously combined hybrid character-level learning with attention mechanisms. Character-level models delve into the text data at a granular level, learning patterns from sequences of characters rather than words, thereby capturing subtle linguistic cues, typos, or uncommon expressions often exploited by deceptive news. This fine-grained analysis is particularly significant in mitigating adversarial attacks that manipulate text to evade detection.

Complementing this, the attention mechanisms act as a cognitive filter, allowing the model to selectively focus on the most relevant parts of an input sequence. By simulating a form of machine cognition akin to human selective attention, this method enables the system to weigh the importance of various textual features dynamically. Such selective emphasis dramatically improves detection precision, particularly in complex, context-rich news articles where the truth might be concealed within elaborate narratives.

Operationalizing this hybrid setup in a distributed federated environment introduces substantial complexity, demanding sophisticated synchronization and optimization strategies. The novel framework expertly addresses these challenges by implementing an efficient communication protocol that minimizes data transfer overhead, thereby maintaining network scalability and reducing operational costs. This approach makes the framework deployable across resource-constrained devices, ranging from smartphones to edge servers, democratizing access to cutting-edge fake news detection technologies.

In practice, the framework continuously refines its detection capabilities by learning from freshly aggregated local model updates contributed by an ever-expanding network of nodes. This iterative process not only enhances model generalizability but also ensures adaptability in the face of evolving fake news tactics. The dynamic feedback-driven learning pipeline equips the system to recognize novel misinformation patterns, making it resilient against the relentless innovation of fake news creators.

Benchmarking experiments conducted by Nithya and Dhivyaa demonstrate the framework’s superior performance on diverse datasets spanning multiple languages and domains. Compared to traditional centralized detection models, this hybrid federated approach consistently exhibits higher accuracy, reduced false positives, and faster convergence rates. Moreover, the computational cost analysis confirms substantial reductions in energy and resource consumption, highlighting its sustainability for real-world deployment at scale.

One of the remarkable advantages of this federated framework is its intrinsic support for privacy-preserving fake news detection across geopolitical boundaries. In an age where national data sovereignty regulations often impede cross-border data sharing, this decentralized learning infrastructure facilitates collective intelligence without compromising compliance. News organizations, governments, and social media platforms across different regions can collaboratively combat misinformation while respecting local privacy laws.

Beyond merely identifying deceitful news, the model’s output provides interpretable insights into the linguistic and semantic elements triggering each detection. This transparency is critical in fostering user trust and enables human moderators to understand and validate automated decisions. By elucidating the underlying rationale, the approach encourages wider adoption among policymakers, content curators, and even everyday users wary of blindly trusting AI verdicts.

Looking ahead, the research opens new frontiers for integrating multimodal data sources into the federated framework. Future expansions could include analyzing images, videos, and audio content alongside textual data to build a more holistic and robust fake news deterrent. The synergistic combination of diverse media inputs with advanced federated learning holds the potential to revolutionize digital media credibility assessment on a global scale.

Furthermore, the researchers emphasize the role of collaborative ecosystem-building among academic institutions, industry, and civil society to continuously refine and sustain such powerful detection mechanisms. Federated deep learning frameworks, especially those championing hybrid character-level and attention-enhanced models, could serve as foundational pillars in building trusted digital information networks that empower users with reliable knowledge.

This landmark study by Nithya and Dhivyaa not only delivers a technological triumph but also a socially impactful tool that addresses one of the most urgent crises of the digital age: the proliferation of fake news. By seamlessly weaving together privacy-preserving federated learning, granular character-level analytics, and context-aware attention processes, their framework exemplifies the confluence of innovation and responsibility in contemporary AI research.

With misinformation continuing to undermine democratic discourse, breed social unrest, and erode trust in institutions, deploying scalable, cost-efficient, and privacy-conscious detection systems is imperative. The federated deep learning methodology illuminated in this research represents a beacon of hope—ushering in an era where AI-powered guardians vigilantly uphold the integrity of information and shield societies from the corrosive effects of falsehood.

In sum, this cutting-edge federated deep learning framework marks a transformative leap toward sustainable, effective, and democratic fake news detection. It paves the way for widespread adoption across diverse platforms and geographies, setting new standards for accuracy, privacy, scalability, and economic viability. As digital communication continues to evolve rapidly, innovations like this stand poised to fortify the frontline defenses against the relentless tide of misinformation threatening the fabric of informed society.

Subject of Research: Fake news detection using federated deep learning frameworks integrating hybrid character-level and attention mechanisms.

Article Title: A federated deep learning framework with distributed hybrid character-level and attention mechanisms for scalable and cost-efficient fake news detection

Article References:
Nithya, K., Dhivyaa, C.R. A federated deep learning framework with distributed hybrid character-level and attention mechanisms for scalable and cost-efficient fake news detection. Sci Rep (2026). https://doi.org/10.1038/s41598-026-54820-6

Image Credits: AI Generated

Federated learning, a decentralized machine learning paradigm, has garnered significant attention for its unique ability to train models across multiple devices or institutions without sharing raw data. While this preserves data confidentiality to an extent, vulnerabilities persist in the communication channels and intermediate data exchanges. Addressing these vulnerabilities, the researchers developed an innovative system that layers stochastic Poisson embedding into the federated learning process. This embedding operates by probabilistically masking user data, adding a buffer of uncertainty to thwart any adversarial inference attempts, thereby augmenting the foundational privacy guarantees.

Central to this framework is the integration of secure homomorphic encryption — a cryptographic marvel that permits computation on encrypted data without decryption. This capability is transformative in scenarios where medical data privacy cannot be compromised. By embedding this encryption strategy into federated learning, the framework allows encrypted gradients and model updates to be processed and aggregated securely, ensuring that no unencrypted sensitive information ever leaves the local data nodes. This not only preserves patient confidentiality but also enables collaborative model improvements over disparate healthcare datasets.

The employment of the stochastic Poisson process introduces a novel randomness element that seamlessly blends with the cryptographic layer. The Poisson process, traditionally used to model random events over time, here serves to generate perturbations in the data representation and model parameters. This stochasticity assists in obfuscating patterns that could otherwise be exploited to reconstruct individual patient records from model gradients, a known vulnerability in federated setups. Importantly, the authors meticulously balance the trade-off between privacy amplification and the accuracy of the resulting AI models.

Medical AI applications stand to benefit immensely from this research because patient data is notoriously sensitive and protected by stringent regulations such as HIPAA and GDPR. Conventional data-sharing constraints often stifle the development of robust machine learning models due to limited access to comprehensive datasets. With the proposed framework, medical institutions can collaboratively develop AI models without exposing raw data, thereby fueling advancements in diagnostics, personalized medicine, and health outcome predictions while steadfastly honoring privacy mandates.

The paper delves deeply into the mathematical formulations underpinning the stochastic Poisson embedding strategy. It elaborates on how the randomization parameters are calibrated dynamically based on the dataset’s distribution characteristics and the model’s iterative update scheme. This adaptability ensures that the introduced noise does not deteriorate model performance, a critical aspect in healthcare settings where accuracy is non-negotiable. By tuning the Poisson process parameters, the algorithm attains a tunable privacy budget, enabling bespoke solutions tailored to institutional risk appetites and regulatory environments.

Furthermore, the homomorphic encryption employed is based on lattice-based cryptosystems, which are resilient to attacks even in the face of emerging quantum computing threats. The authors detail the selection and optimization of encryption parameters that achieve a pragmatic balance, offering computational efficiency without compromising security. The encryption overhead, a common drawback in homomorphic encryption schemes, is mitigated through algorithmic optimizations and parallel processing techniques, rendering the framework viable for real-world medical federated learning deployments.

An extensive experimental evaluation demonstrates the efficacy of this method across multiple medical datasets, including imaging and electronic health records. Notably, the framework maintains high model accuracy, with negligible degradation compared to non-private federated learning baselines. Simultaneously, privacy leakage metrics indicate a substantial decrease in potential information exposure, underscoring the robustness of the proposed solution. These results are crucial in establishing the operational viability of privacy-preserving AI in clinical contexts.

The research team also addresses potential adversarial models, considering both honest-but-curious participants who follow protocol but attempt to infer private data, and malicious insiders who might deviate from prescribed behaviors. Their framework includes safeguards such as anomaly detection modules that monitor gradient updates for signs of abnormal data manipulation. These multi-layered defense mechanisms provide a comprehensive security posture that is essential when handling high-stakes healthcare data.

Importantly, the framework is generalizable beyond medical AI and can be adapted to any federated learning setting where privacy is of utmost concern. The inclusion of a probabilistic embedding combined with homomorphic encryption serves as a blueprint for future privacy frameworks seeking to tame the privacy-performance paradox in distributed machine learning environments. This cross-domain potential vastly amplifies the impact and applicability of the findings presented in this study.

From an ethical perspective, this work contributes to restoring patient trust in AI-based medical technologies by ensuring that privacy is baked into the design of data collaborations rather than retrofitted afterward. Given the increasing prevalence of data breaches and privacy scandals, such privacy-first approaches will likely become a mandatory standard. The research also supports compliance with evolving legal frameworks, potentially easing the bureaucratic hurdles associated with cross-institutional data sharing and AI model development.

Technologically, implementing the stochastic Poisson-embedded privacy framework demands sophisticated infrastructure capable of running encrypted computations and managing dynamic perturbations efficiently. The authors suggest leveraging modern cloud-based solutions with secure enclaves and specialized hardware accelerators, which could catalyze wider adoption by lowering technical entry barriers. Future research directions proposed include exploring adaptive privacy budgets controlled by continuous risk assessment and enhancing the framework’s scalability for ultra-large federated networks.

In conclusion, the study authored by Gomathi et al. represents a significant leap in privacy-aware federated learning tailored for the medical AI domain. By marrying stochastic statistical processes with cutting-edge cryptographic techniques, the framework sets a new benchmark for secure collaborative intelligence, empowering stakeholders to innovate without sacrificing the sanctity of personal health information. This advancement is poised to accelerate the integration of AI into healthcare, promising both technological progress and enhanced patient privacy.

Subject of Research:

Article Title:

Article References:

Gomathi, R., Saranya, K., Mahaboob John, Y.M. et al. Stochastic Poisson-embedded privacy framework for federated learning with secure homomorphic encryption in medical AI.
Sci Rep (2026). https://doi.org/10.1038/s41598-026-41469-4

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s41598-026-41469-4

Keywords: