Federated learning(FL)is a distributed machine learning paradigm for edge cloud computing.FL can facilitate data-driven decision-making in tactical scenarios,effectively addressing both data volume and infrastructure ...Federated learning(FL)is a distributed machine learning paradigm for edge cloud computing.FL can facilitate data-driven decision-making in tactical scenarios,effectively addressing both data volume and infrastructure challenges in edge environments.However,the diversity of clients in edge cloud computing presents significant challenges for FL.Personalized federated learning(pFL)received considerable attention in recent years.One example of pFL involves exploiting the global and local information in the local model.Current pFL algorithms experience limitations such as slow convergence speed,catastrophic forgetting,and poor performance in complex tasks,which still have significant shortcomings compared to the centralized learning.To achieve high pFL performance,we propose FedCLCC:Federated Contrastive Learning and Conditional Computing.The core of FedCLCC is the use of contrastive learning and conditional computing.Contrastive learning determines the feature representation similarity to adjust the local model.Conditional computing separates the global and local information and feeds it to their corresponding heads for global and local handling.Our comprehensive experiments demonstrate that FedCLCC outperforms other state-of-the-art FL algorithms.展开更多
This letter provides a critical appraisal of the comprehensive meta-analysis by Hou et al,which synthesizes the incidence and risk factors for postoperative delirium(POD)in organ transplant recipients.Their work estab...This letter provides a critical appraisal of the comprehensive meta-analysis by Hou et al,which synthesizes the incidence and risk factors for postoperative delirium(POD)in organ transplant recipients.Their work establishes a pooled POD incidence of 20%,with significant variability across organ types(lung 34%,liver 22%,kidney 6%),and identifies key risk factors including primary graft dysfunction,hepatic encephalopathy,and high model for end-stage liver disease/acute physiology and chronic health evaluation Ⅱ scores.This commentary acknowledges the study's strength in providing a robust,trans-organ synthesis of current evidence.However,it critically discusses the substantial heterogeneity,the counterintuitive non-significance of age as a risk factor,and the unavoidable limitation of unmeasured confounders inherent in meta-analyses,such as preoperative cognitive/psychiatric status and anesthetic protocols.While the findings provide an essential evidence base for risk stratification and prevention,this letter argues that the high heterogeneity underscores the need for organ-specific analysis and calls for large-scale,prospective studies with standardized protocols to translate these findings into reliable clinical prediction tools and targeted interventions.展开更多
基金supported by the Natural Science Foundation of Xinjiang Uygur Autonomous Region(Grant No.2022D01B 187)。
文摘Federated learning(FL)is a distributed machine learning paradigm for edge cloud computing.FL can facilitate data-driven decision-making in tactical scenarios,effectively addressing both data volume and infrastructure challenges in edge environments.However,the diversity of clients in edge cloud computing presents significant challenges for FL.Personalized federated learning(pFL)received considerable attention in recent years.One example of pFL involves exploiting the global and local information in the local model.Current pFL algorithms experience limitations such as slow convergence speed,catastrophic forgetting,and poor performance in complex tasks,which still have significant shortcomings compared to the centralized learning.To achieve high pFL performance,we propose FedCLCC:Federated Contrastive Learning and Conditional Computing.The core of FedCLCC is the use of contrastive learning and conditional computing.Contrastive learning determines the feature representation similarity to adjust the local model.Conditional computing separates the global and local information and feeds it to their corresponding heads for global and local handling.Our comprehensive experiments demonstrate that FedCLCC outperforms other state-of-the-art FL algorithms.
基金Supported by National Research Foundation of Korea,No.RS-2023-00237287.
文摘This letter provides a critical appraisal of the comprehensive meta-analysis by Hou et al,which synthesizes the incidence and risk factors for postoperative delirium(POD)in organ transplant recipients.Their work establishes a pooled POD incidence of 20%,with significant variability across organ types(lung 34%,liver 22%,kidney 6%),and identifies key risk factors including primary graft dysfunction,hepatic encephalopathy,and high model for end-stage liver disease/acute physiology and chronic health evaluation Ⅱ scores.This commentary acknowledges the study's strength in providing a robust,trans-organ synthesis of current evidence.However,it critically discusses the substantial heterogeneity,the counterintuitive non-significance of age as a risk factor,and the unavoidable limitation of unmeasured confounders inherent in meta-analyses,such as preoperative cognitive/psychiatric status and anesthetic protocols.While the findings provide an essential evidence base for risk stratification and prevention,this letter argues that the high heterogeneity underscores the need for organ-specific analysis and calls for large-scale,prospective studies with standardized protocols to translate these findings into reliable clinical prediction tools and targeted interventions.
