Local Model Research Articles

Many-body localized hidden generative models

Quantum generative models hold the promise of accelerating or improving machine learning tasks by leveraging the probabilistic nature of quantum states, but the successful optimization of these models remains a difficult challenge. To tackle this challenge, we present a new architecture for quantum generative modeling that combines insights from classical machine learning and quantum phases of matter. In particular, our model utilizes both many-body localized (MBL) dynamics and hidden units to improve the optimization of the model. We demonstrate the applicability of our model on a diverse set of classical and quantum tasks, including a toy version of MNIST handwritten digits, quantum data obtained from quantum many-body states, and nonlocal parity data. Our architecture and algorithm provide novel strategies of utilizing quantum many-body systems as learning resources and reveal a powerful connection between disorder, interaction, and learning in quantum many-body systems. Published by the American Physical Society 2024

To be or not to be, when synthetic data meet clinical pharmacology: A focused study on pharmacogenetics.

The use of synthetic data in pharmacology research has gained significant attention due to its potential to address privacy concerns and promote open science. In this study, we implemented and compared three synthetic data generation methods, CT-GAN, TVAE, and a simplified implementation of Avatar, for a previously published pharmacogenetic dataset of 253 patients with one measurement per patient (non-longitudinal). The aim of this study was to evaluate the performance of these methods in terms of data utility and privacy trade off. Our results showed that CT-GAN and Avatar used with k = 10 (number of patients used to create the local model of generation) had the best overall performance in terms of data utility and privacy preservation. However, the TVAE method showed a relatively lower level of performance in these aspects. In terms of Hazard ratio estimation, Avatar with k = 10 produced HR estimates closest to the original data, whereas CT-GAN slightly underestimated the HR and TVAE showed the most significant deviation from the original HR. We also investigated the effect of applying the algorithms multiple times to improve results stability in terms of HR estimation. Our findings suggested that this approach could be beneficial, especially in the case of small datasets, to achieve more reliable and robust results. In conclusion, our study provides valuable insights into the performance of CT-GAN, TVAE, and Avatar methods for synthetic data generation in pharmacogenetic research. The application to other type of data and analyses (data driven) used in pharmacology should be further investigated.

Geographically weighted accelerated failure time model for spatial survival data: application to ovarian cancer survival data in New Jersey

BackgroundIn large multiregional cohort studies, survival data is often collected at small geographical levels (such as counties) and aggregated at larger levels, leading to correlated patterns that are associated with location. Traditional studies typically analyze such data globally or locally by region, often neglecting the spatial information inherent in the data, which can introduce bias in effect estimates and potentially reduce statistical power.MethodWe propose a Geographically Weighted Accelerated Failure Time Model for spatial survival data to investigate spatial heterogeneity. We establish a weighting scheme and bandwidth selection based on quasi-likelihood information criteria. Theoretical properties of the proposed estimators are thoroughly examined. To demonstrate the efficacy of the model in various scenarios, we conduct a simulation study with different sample sizes and adherence to the proportional hazards assumption or not. Additionally, we apply the proposed method to analyze ovarian cancer survival data from the Surveillance, Epidemiology, and End Results cancer registry in the state of New Jersey.ResultsOur simulation results indicate that the proposed model exhibits superior performance in terms of four measurements compared to existing methods, including the geographically weighted Cox model, when the proportional hazards assumption is violated. Furthermore, in scenarios where the sample size per location is 20-25, the simulation data failed to fit the local model, while our proposed model still demonstrates satisfactory performance. In the empirical study, we identify clear spatial variations in the effects of all three covariates.ConclusionOur proposed model offers a novel approach to exploring spatial heterogeneity of survival data compared to global and local models, providing an alternative to geographically weighted Cox regression when the proportional hazards assumption is not met. It addresses the issue of certain counties' survival data being unable to fit the model due to limited samples, particularly in the context of rare diseases.

Data Quality-Aware Client Selection in Heterogeneous Federated Learning

Federated Learning (FL) enables decentralized data utilization while maintaining edge user privacy, but it faces challenges due to statistical heterogeneity. Existing approaches address client drift and data heterogeneity issues. However, real-world settings often involve low-quality data with noisy features, such as covariate drift or adversarial samples, which are usually ignored. Noisy samples significantly impact the global model’s accuracy and convergence rate. Assessing data quality and selectively aggregating updates from high-quality clients is crucial, but dynamically perceiving data quality without additional computations or data exchanges is challenging. In this paper, we introduce the FedDQA (Federated learning via Data Quality-Aware) (FedDQA) framework. We discover increased data noise leads to slower loss reduction during local model training. We propose a loss sharpness-based Data-Quality-Awareness (DQA) metric to differentiate between high-quality and low-quality data. Based on the DQA, we design a client selection algorithm that strategically selects participant clients to reduce the negative impact of noisy clients. Experiment results indicate that FedDQA significantly outperforms the baselines. Notably, it achieves up to a 4% increase in global model accuracy and demonstrates faster convergence rates.

MedLSAM: Localize and segment anything model for 3D CT images

Recent advancements in foundation models have shown significant potential in medical image analysis. However, there is still a gap in models specifically designed for medical image localization. To address this, we introduce MedLAM, a 3D medical foundation localization model that accurately identifies any anatomical part within the body using only a few template scans. MedLAM employs two self-supervision tasks: unified anatomical mapping (UAM) and multi-scale similarity (MSS) across a comprehensive dataset of 14,012 CT scans. Furthermore, we developed MedLSAM by integrating MedLAM with the Segment Anything Model (SAM). This innovative framework requires extreme point annotations across three directions on several templates to enable MedLAM to locate the target anatomical structure in the image, with SAM performing the segmentation. It significantly reduces the amount of manual annotation required by SAM in 3D medical imaging scenarios. We conducted extensive experiments on two 3D datasets covering 38 distinct organs. Our findings are twofold: (1) MedLAM can directly localize anatomical structures using just a few template scans, achieving performance comparable to fully supervised models; (2) MedLSAM closely matches the performance of SAM and its specialized medical adaptations with manual prompts, while minimizing the need for extensive point annotations across the entire dataset. Moreover, MedLAM has the potential to be seamlessly integrated with future 3D SAM models, paving the way for enhanced segmentation performance. Our code is public at https://github.com/openmedlab/MedLSAM.

Caries Detection and Classification in Photographs Using an Artificial Intelligence-Based Model-An External Validation Study.

This ex vivo diagnostic study aimed to externally validate a freely accessible AI-based model for caries detection, classification, localisation and segmentation using an independent image dataset. It was hypothesised that there would be no difference in diagnostic performance compared to previously published internal validation data. For the independent dataset, 718 dental images representing different stages of carious (n = 535) and noncarious teeth (n = 183) were retrieved from the internet. All photographs were evaluated by the dental team (reference standard) and the AI-based model (test method). Diagnostic performance was statistically determined using cross-tabulations to calculate accuracy (ACC), sensitivity (SE), specificity (SP) and area under the curve (AUC). An overall ACC of 92.0% was achieved for caries detection, with an ACC of 85.5-95.6%, SE of 42.9-93.3%, SP of 82.1-99.4% and AUC of 0.702-0.909 for the classification of caries. Furthermore, 97.0% of the cases were accurately localised. Fully and partially correct segmentation was achieved in 52.9% and 44.1% of the cases, respectively. The validated AI-based model showed promising diagnostic performance in detecting and classifying caries using an independent image dataset. Future studies are needed to investigate the validity, reliability and practicability of AI-based models using dental photographs from different image sources and/or patient groups.

Generalized juggling patterns, quiver Grassmannians and affine flag varieties

The goal of this paper is to clarify the connection between certain structures from the theory of totally nonnegative Grassmannians, quiver Grassmannians for cyclic quivers and the theory of local models of Shimura varieties. More precisely, we generalize the construction from our previous paper relating the combinatorics and geometry of quiver Grassmannians to that of the totally nonnegative Grassmannians. The varieties we are interested in serve as realizations of local models of Shimura varieties. We exploit quiver representation techniques to study the quiver Grassmannians of interest and, in particular, to describe explicitly embeddings into affine flag varieties which allow us to realize our quiver Grassmannians as a union of Schubert varieties therein.

Enhancing Heart Disease Prediction with Federated Learning and Blockchain Integration

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

Bolt loosening assessment using ensemble vision models for automatic localization and feature extraction with target‐free perspective adaptation

AbstractBolt loosening assessment is crucial to identify early warnings of structural degradation and prevent catastrophic events. This paper proposes an automatic bolt loosening assessment methodology. First, a novel end‐to‐end ensemble vision model, Bolt‐FP‐Net, is proposed to reason the locations of bolts and their hexagonal feature patterns concurrently. Second, an adaptive target‐free perspective correction method is proposed to correct perspective distortion and enhance assessment accuracy. Finally, an iterative bolt loosening quantification is developed to estimate and refine the bolt loosening rotation. Experimental parametric studies indicated that the proposed Bolt‐FP‐Net can achieve excellent performance under different environmental conditions. Finally, a case study was conducted on steel bolt connections, which shows the proposed methodology can achieve high accuracy and real‐time speed in bolt loosening assessment.

Shear-driven magnetic buoyancy in the solar tachocline: Dependence of the mean electromotive force on diffusivity and latitude

Abstract The details of the dynamo process that is responsible for driving the solar magnetic activity cycle are still not fully understood. In particular, whilst differential rotation provides a plausible mechanism for the regeneration of the toroidal (azimuthal) component of the large-scale magnetic field, there is ongoing debate regarding the process that is responsible for regenerating the Sun’s large-scale poloidal field. Our aim is to demonstrate that magnetic buoyancy, in the presence of rotation, is capable of producing the necessary regenerative effect. Building upon our previous work, we carry out numerical simulations of a local Cartesian model of the tachocline, consisting of a rotating, fully compressible, electrically conducting fluid with a forced shear flow. An initially weak, vertical magnetic field is sheared into a strong, horizontal magnetic layer that becomes subject to magnetic buoyancy instability. By increasing the Prandtl number we lessen the back reaction of the Lorentz force onto the shear flow, maintaining stronger shear and a more intense magnetic layer. This in turn leads to a more vigorous instability and a much stronger mean electromotive force, which has the potential to significantly influence the evolution of the mean magnetic field. These results are only weakly dependent upon the inclination of the rotation vector, i.e. the latitude of the local Cartesian model. Although further work is needed to confirm this, these results suggest that magnetic buoyancy in the tachocline is a viable poloidal field regeneration mechanism for the solar dynamo.

Comparative validation of global and climate-specific decomposition models at high latitudes

Reliable modelling of the solar potential of urban surfaces (i.e. roofs, facades and ground) in the built environment can contribute to boost the exploitation of solar energy. Empirical and quasi-physical models are combined into model chains for this scope. Among these models, the decomposition models are used to separate the direct from the diffuse solar irradiance. Such variables are rarely measured, thus making necessary their computation. Studies in literature pointed out that Yang4 is the best-performing decomposition model globally. However, in geographically limited applications, quasi-universal decomposition models like Yang4 and Engerer4 can be outperformed by local models (i.e. models parametrized with climate-specific data) such as Skartveit3 and Starke3. This makes necessary to perform local validation studies to verify the findings from worldwide validation studies. In this study, the four best-performing decomposition models are implemented in Python and experimentally validated against 1-min solar irradiance data (i.e. direct and diffuse irradiance) of Trondheim (Lat. 63°26′N, Norway). More than 1-year long data are considered. The study confirms that the Yang4 model performed the best for high-latitude applications, but Starke3 can achieve a level of performance close to Yang4.

Improving a shell-tube latent heat thermal energy storage unit for building hot water demand using metal foam inserts at a constant pumping power

The phase transition heat transfer during the melting and solidification processes of phase change materials (PCMs) was modeled in a shell-tube thermal energy storage unit. For the first time, special attention was dedicated to the heat transfer enhancement on the heat transfer fluid (HTF) side by placing metal foam (MF) inserts in the HTF tube. The MF inserts on the HTF side were placed next to the fin bases to produce the maximum heat transfer enhancement. A two-temperature local thermal non-equilibrium model was applied to accurately model the transient heat transfer in the MF domains. To perform a fair heat transfer analysis and consider the hydrodynamic aspects of heat transfer, the pumping power for driving the HTF fluid was kept fixed. The finite element method with automatic time-step control was used to integrate the model equations. Results show that the insertion of the MF in the HTF tube enhances conduction heat transfer but reduces flow for a constant pump power. Using a small amount of MF provides a good flow rate and a convective heat transfer while using a large amount of MF provides good conduction heat transfer but a poor flow rate and convection. A moderate amount of MF inserts results in poor conduction and a poor flow rate, notably increasing the phase transition time. A 5 % MF insert could provide about a 13 % shorter solidification time compared to a 25 % MF insert. Thus, a well-designed and engineered MF insert configuration on the HTF side can reduce the phase transition time and improve the energy storage power.

Towards value-sensitive and poisoning-proof model aggregation for federated learning on heterogeneous data

Federated Learning (FL) enables collaborative model training without sharing data, but traditional static averaging of local updates leads to poor performance on heterogeneous data. The following remedies, either by scheduling data distribution or mitigating local discrepancies, predominately fail to handle fine-grained heterogeneity (e.g., local imbalanced labels). To commence, we reveal that static averaging leads to the global model suffering from the mean fallacy. That is, the averaging process favors the local model with large parameters numerically rather than knowledge. To tackle this, we introduce FedVSA, a simple-yet-effective model aggregation framework sensitive to heterogeneous local data merits. Specifically, we invent a new global loss function for FL by prioritizing the valuable local updates, facilitating efficient convergence. We deduce a softmax-based aggregation rule and prove its convergence property via rigorous theoretical analysis. Additionally, we expose poisoning threats of model replacement that utilize the mean fallacy for attacks. To mitigate this threat, we propose a two-step mechanism involving auditing historic local training statistics and analyzing the Shapley Value. Through extensive experiments, we show that FedVSA achieves faster convergence (~1.52×) and higher accuracy (~1.6%) compared to the baselines. It also effectively mitigates poisoning attacks by agilely recovering and returning to normal aggregation.

Signal-receiving UAV Positioning Model Based on Geometric Analysis

A drone is an aerial vehicle without a pilot or controller that is remotely controlled using aerodynamics to fulfil a designated task under certain circumstances. With the encouragement and support of national policies, the development of China's UAV industry has a good prospect. In real life, location information is the key information for UAVs to accurately complete operations and realise mission requirements, and the importance of its accurate positioning is self-evident. For the UAV localisation model, the bias-free localisation model is established first, and then the localisation model in the case of bias is solved on this basis. When there is a deviation in the received signal UAV, the UAV may be located outside the circle, on the circle, and inside the circle. Specifically divided into four scenarios, see Figure 1-4. (Analysed: outside the circle: θ1’<θ1; inside the circle: θ1’>θ1; circle on the left: θ2>θ2’, right: θ2<θ2’; no deviation: θ1’=θ1, θ2’<θ2). When the position of the passive receiving signal UAV does not exist deviation, according to the position relationship obtained four cases, respectively, the establishment of the quadrilateral, the use of sinusoidal theorem to obtain the receiving signal UAV positioning model, the same can be obtained under the existence of deviation, the specific positioning model of the receiving signal UAV.

BRFL: A blockchain-based byzantine-robust federated learning model

With the increasing importance of machine learning, the privacy and security of training data have become a concern. Federated learning, which stores data in distributed nodes and shares only model parameters, has gained significant attention for addressing this concern. However, a challenge arises in federated learning due to the byzantine attack problem, where malicious local models can compromise the global model's performance during aggregation. This article proposes the Blockchain-based Byzantine-Robust Federated Learning (BRFL) model, which combines federated learning with blockchain technology. We improve the robustness of federated learning by proposing a new consensus algorithm and aggregation algorithm for blockchain-based federated learning. Meanwhile, we modify the block saving rules of the blockchain to reduce the storage pressure of the nodes. Experimental results on public datasets demonstrate the superior byzantine robustness of our secure aggregation algorithm compared to other baseline aggregation methods, and reduce the storage pressure of the blockchain nodes.

Distributed Fire Classification and Localization Model Based on Federated Learning with Image Clustering

Distributed Fire Classification and Localization Model Based on Federated Learning with Image Clustering

PQSF: post-quantum secure privacy-preserving federated learning

In federated learning, secret sharing is a key technology to maintain the privacy of participants’ local models. Moreover, with the rapid development of quantum computers, existing federated learning privacy protection schemes based on secret sharing will no longer be able to guarantee the data security of participants in the post-quantum era. In addition, existing privacy protection methods have the problem of high communication and computational overhead. Although the multi-stage secret sharing scheme proposed by Pilaram et al. is one of the effective solutions to the above problems, existing studies have proven the privacy leakage risk of this scheme. This paper firstly designs a new lattice-based multi-stage secret sharing scheme Improved-Pilaram to solve the security problem, which allows participants to use public vectors to reconstruct different secret values without changing the secret sharing. Based on Improved-Pilaram, this article proposes a post-quantum secure federated learning scheme PQSF. PQSF uses double masking technology to encrypt model parameters and achieves mask reconstruction through secret sharing. Since Improved-Pilaram is multi-stage, participants do not need to update their local secret shares frequently during training. Analysis and experimental results show that the PQSF proposed in this paper reduces the communication complexity between participants and reduces the computational overhead by about 20% compared with existing solutions.

A Weibo local network growth model constructed from the perspective of following-followed

Abstract In order to explore the evolution process of the Weibo local network, this study first defines four factors influencing the evolution of the Weibo network. On this basis, the BA scale-free network model was enhanced by incorporating these four factors and accounting for directionality, resulting in a Weibo local network evolution model based on user attributes and behavioral similarity. The model's validity was validated by comparing simulation results with real data. The findings indicate that the Weibo local network exhibits both small-world characteristics and distinctive features. The results show that the Weibo local network exhibits both small-world characteristics and distinctive properties. The in-degree distribution follows a mixed pattern of exponential and power-law distributions, the degree-degree shows isomatching, and both the in-degree centrality and eigenvector centrality values are relatively low. This research contributes to our understanding of user behaviour in the Weibo network, and provides a structural basis for exploring the impact of Weibo network structure on information dissemination.

A Federated Transfer Fault Diagnosis Method for Cross-Domain and Incomplete Data

In fact, multiple clients often use similar devices and collect fault data separately, so joint multi-client collaborative fault diagnosis modeling can solve the problem of data scarcity, but this poses great challenges to data privacy protection. In this paper, we propose a federated transfer fault diagnosis method based on federated learning for cross-domain incomplete data. The proposed method only exchanges the parameters of the local training model, which achieves the privacy protection of the client’s local data. We construct a multi-client collaborative learning framework to address the problem of weak generalization ability caused by the lack of terms in single client training samples. We also propose a targeted semi-supervised fine-tuning strategy based on relative distance to reduce the probability of negative fine-tuning of out-of-distribution samples and improve the accuracy of diagnostic models. The results of cross-condition and cross-equipment experiments demonstrate that the proposed method has obvious advantages over the existing fault diagnosis methods.

A blueprint for coupling a hydrological model with fine- and coarse-scale atmospheric regional climate change models for probabilistic streamflow projections

In cold regions, climate change, including warming and changes in snowmelt dynamics, have profound impacts on streamflow patterns, often leading to flooding events. Understanding the projected changes in hydrometeorological factors contributing to streamflow, such as snowmelt runoff and rain-on-snowmelt, is crucial for effective adaptation and mitigation strategies. It is also essential to consider uncertainty in streamflow projections and incorporate this uncertainty into flood predictions. This study presents a blueprint for calculating probabilistic future streamflow and flow duration curves in a mountainous cold region. The methodology integrated forcings from an atmospheric model (WRF) with a hydrological model (MESH) at fine spatial (4 km) and temporal (3-hourly) resolutions. To account for uncertainty, an ensemble of 15 CanRCM4 regional climate simulations with varying boundary conditions for the RCP 8.5 scenario was utilized. A novel method was developed to perturb the CanRCM4 simulations’ precipitation using a WRF Pseudo Global Warming run to incorporate uncertainty. This comprehensive approach revealed significant uncertainty in streamflow driven by internal variability. WRF-driven simulations without uncertainty showed a decrease in future streamflow extremes, while the perturbed simulations demonstrated the potential for substantially higher values under climate change. Moreover, the study derived probabilistic flow duration curves from the streamflow projections, aiding in estimating flood frequency for floodplain mapping when driving local hydraulic models. The methodology developed in this work can be extended to other river basins in Canada and elsewhere where gridded downscaled climate model outputs are available.