With the rapid development of mobile internet technology and increasing concerns over data privacy,Federated Learning(FL)has emerged as a significant framework for training machine learning models.Given the advancemen...With the rapid development of mobile internet technology and increasing concerns over data privacy,Federated Learning(FL)has emerged as a significant framework for training machine learning models.Given the advancements in technology,User Equipment(UE)can now process multiple computing tasks simultaneously,and since UEs can have multiple data sources that are suitable for various FL tasks,multiple tasks FL could be a promising way to respond to different application requests at the same time.However,running multiple FL tasks simultaneously could lead to a strain on the device’s computation resource and excessive energy consumption,especially the issue of energy consumption challenge.Due to factors such as limited battery capacity and device heterogeneity,UE may fail to efficiently complete the local training task,and some of them may become stragglers with high-quality data.Aiming at alleviating the energy consumption challenge in a multi-task FL environment,we design an automatic Multi-Task FL Deployment(MFLD)algorithm to reach the local balancing and energy consumption goals.The MFLD algorithm leverages Deep Reinforcement Learning(DRL)techniques to automatically select UEs and allocate the computation resources according to the task requirement.Extensive experiments validate our proposed approach and showed significant improvements in task deployment success rate and energy consumption cost.展开更多
Background:Breast cancer is one of the most common cancer in women and a proportion of patients experiences brain metastases with poor prognosis.The study aimed to construct a novel predictive clinical model to evalua...Background:Breast cancer is one of the most common cancer in women and a proportion of patients experiences brain metastases with poor prognosis.The study aimed to construct a novel predictive clinical model to evaluate the overall survival(OS)of patients with postoperative brain metastasis of breast cancer(BCBM)and validate its effectiveness.Methods:From 2010 to 2020,a total of 310 female patients with BCBM were diagnosed in The Affiliated Cancer Hospital of Xinjiang Medical University,and they were randomly assigned to the training cohort and the validation cohort.Data of another 173 BCBM patients were collected from the Surveillance,Epidemiology,and End Results Program(SEER)database as an external validation cohort.In the training cohort,the least absolute shrinkage and selection operator(LASSO)Cox regression model was used to determine the fundamental clinical predictive indicators and the nomogram was constructed to predict OS.The model capability was assessed using receiver operating characteristic,C-index,and calibration curves.Kaplan-Meier survival analysis was performed to evaluate clinical effectiveness of the risk stratification system in the model.The accuracy and prediction capability of the model were verified using the validation and SEER cohorts.Results:LASSO Cox regression analysis revealed that lymph node metastasis,molecular subtype,tumor size,chemotherapy,radiotherapy,and lung metastasis were statistically significantly correlated with BCBM.The C-indexes of the survival nomogram in the training,validation,and SEER cohorts were 0.714,0.710,and 0.670,respectively,which showed good prediction capability.The calibration curves demonstrated that the nomogram had great forecast precision,and a dynamic diagram was drawn to increase the maneuverability of the results.The Risk Stratification System showed that the OS of lowrisk patients was considerably better than that of high-risk patients(P<0.001).Conclusion:The nomogram prediction model constructed in this study has a good predictive value,which can effectively evaluate the survival rate of patients with postoperative BCBM.展开更多
文摘With the rapid development of mobile internet technology and increasing concerns over data privacy,Federated Learning(FL)has emerged as a significant framework for training machine learning models.Given the advancements in technology,User Equipment(UE)can now process multiple computing tasks simultaneously,and since UEs can have multiple data sources that are suitable for various FL tasks,multiple tasks FL could be a promising way to respond to different application requests at the same time.However,running multiple FL tasks simultaneously could lead to a strain on the device’s computation resource and excessive energy consumption,especially the issue of energy consumption challenge.Due to factors such as limited battery capacity and device heterogeneity,UE may fail to efficiently complete the local training task,and some of them may become stragglers with high-quality data.Aiming at alleviating the energy consumption challenge in a multi-task FL environment,we design an automatic Multi-Task FL Deployment(MFLD)algorithm to reach the local balancing and energy consumption goals.The MFLD algorithm leverages Deep Reinforcement Learning(DRL)techniques to automatically select UEs and allocate the computation resources according to the task requirement.Extensive experiments validate our proposed approach and showed significant improvements in task deployment success rate and energy consumption cost.
基金supported by National Natural Science Foundation of China(No.82060520)Tianshan Cedar Talent Training Project of Science and Technology Department of Xinjiang Uygur Autonomous Region(No.2020XS14).
文摘Background:Breast cancer is one of the most common cancer in women and a proportion of patients experiences brain metastases with poor prognosis.The study aimed to construct a novel predictive clinical model to evaluate the overall survival(OS)of patients with postoperative brain metastasis of breast cancer(BCBM)and validate its effectiveness.Methods:From 2010 to 2020,a total of 310 female patients with BCBM were diagnosed in The Affiliated Cancer Hospital of Xinjiang Medical University,and they were randomly assigned to the training cohort and the validation cohort.Data of another 173 BCBM patients were collected from the Surveillance,Epidemiology,and End Results Program(SEER)database as an external validation cohort.In the training cohort,the least absolute shrinkage and selection operator(LASSO)Cox regression model was used to determine the fundamental clinical predictive indicators and the nomogram was constructed to predict OS.The model capability was assessed using receiver operating characteristic,C-index,and calibration curves.Kaplan-Meier survival analysis was performed to evaluate clinical effectiveness of the risk stratification system in the model.The accuracy and prediction capability of the model were verified using the validation and SEER cohorts.Results:LASSO Cox regression analysis revealed that lymph node metastasis,molecular subtype,tumor size,chemotherapy,radiotherapy,and lung metastasis were statistically significantly correlated with BCBM.The C-indexes of the survival nomogram in the training,validation,and SEER cohorts were 0.714,0.710,and 0.670,respectively,which showed good prediction capability.The calibration curves demonstrated that the nomogram had great forecast precision,and a dynamic diagram was drawn to increase the maneuverability of the results.The Risk Stratification System showed that the OS of lowrisk patients was considerably better than that of high-risk patients(P<0.001).Conclusion:The nomogram prediction model constructed in this study has a good predictive value,which can effectively evaluate the survival rate of patients with postoperative BCBM.
