In non-independent and identically distributed(non-IID)data environments,model performance often degrades significantly.To address this issue,two improvement methods are proposed:FedReg and FedReg^(*).FedReg is a meth...In non-independent and identically distributed(non-IID)data environments,model performance often degrades significantly.To address this issue,two improvement methods are proposed:FedReg and FedReg^(*).FedReg is a method based on hybrid regularization aimed at enhancing federated learning in non-IID scenarios.It introduces hybrid regularization to replace traditional L2 regularization,combining the advantages of L1 and L2 regularization to enable feature selection while preventing overfitting.This method better adapts to the diverse data distributions of different clients,improving the overall model performance.FedReg^(*)combines hybrid regularization with weighted model aggregation.In addition to the benefits of hybrid regularization,FedReg^(*)applies a weighted averaging method in the model aggregation process,calculating weights based on the cosine similarity between each client gradient and the global gradient to more reasonably distribute client contributions.By considering variations in data quality and quantity among clients,FedReg^(*)highlights the importance of key clients and enhances the model’s generalization performance.These improvement methods enhance model accuracy and communication efficiency.展开更多
联邦学习因具有隐私保护的天然特性,已经逐渐成为一个被广泛认可的分布式机器学习框架。但由于参与方数据分布的差异性,特别是呈现非独立同分布(Non-Independent and Identically Distributed,Non-IID)时,其面临着泛化性能不足、收敛性...联邦学习因具有隐私保护的天然特性,已经逐渐成为一个被广泛认可的分布式机器学习框架。但由于参与方数据分布的差异性,特别是呈现非独立同分布(Non-Independent and Identically Distributed,Non-IID)时,其面临着泛化性能不足、收敛性能下降、数据倾斜等严峻挑战。用预训练基础模型缓解Non-IID问题作为一种新颖的方法,演变出了各种各样的解决方案。对此,从预训练基础模型的角度,对现有工作进行了综述。首先介绍了基础模型方法,对典型的基础模型编码结构进行对比分析。其次从修改输入、基础模型部分结构再训练,以及参数高效微调3个角度,提出了一种新的分类方法。最后探讨了该类工作的核心难题和未来研究方向。展开更多
目的评估血清非高密度脂蛋白胆固醇与高密度脂蛋白胆固醇比值(non-high density lipoprotein cholesterol/high density lipoprotein cholesterol,NHHR)、血浆致动脉粥样硬化指数(atherogenic index of plasma,AIP)与慢性肾病(chronic k...目的评估血清非高密度脂蛋白胆固醇与高密度脂蛋白胆固醇比值(non-high density lipoprotein cholesterol/high density lipoprotein cholesterol,NHHR)、血浆致动脉粥样硬化指数(atherogenic index of plasma,AIP)与慢性肾病(chronic kidney disease,CKD)发病风险的关系,为CKD的防治提供依据。方法以金昌队列中25377名未患CKD的参与者作为研究对象,采用Cox比例风险回归模型、限制性立方样条分别评估NHHR和AIP对CKD的发病风险及剂量-反应关系,并进行亚组分析。采用受试者工作特征曲线评估NHHR和AIP对CKD发病风险的预测能力。结果经过平均4.77年的随访调查后,新发CKD患者有1213例,发病密度为10.03/1000人年。调整混杂因素后,相较于Q1组,Q4组人群中NHHR和AIP的CKD发病风险比分别为1.270(95%CI:1.066~1.512)和1.294(95%CI:1.081~1.548),且均存在一定的剂量-反应关系(均P<0.05)。NHHR和AIP预测CKD的AUC值分别为0.750(95%CI:0.736~0.764)和0.735(95%CI:0.721~0.749)。亚组分析发现,吸烟和糖尿病与NHHR、糖尿病和AIP间存在交互作用(均P<0.05)。结论NHHR和AIP是CKD发病的独立危险因素,并对CKD发病风险有一定的预测能力。展开更多
在客户端数据非独立同分布(Non‑Independent and Identically Distributed,Non‑IID)的场景中,为了向客户端提供个性化且通信高效的解决方案,提出了一种面向Non‑IID场景的通信高效个性化联邦学习算法。具体地,为充分利用相似客户端之间...在客户端数据非独立同分布(Non‑Independent and Identically Distributed,Non‑IID)的场景中,为了向客户端提供个性化且通信高效的解决方案,提出了一种面向Non‑IID场景的通信高效个性化联邦学习算法。具体地,为充分利用相似客户端之间的知识提升模型性能,同时保留本地客户端的个性化信息,提出一种融合模型分层思想与聚类思想的个性化联邦学习算法。为进一步解决通信开销高的问题,设计选择性模型聚合策略,在中心服务器通过最大均值差异评估客户端数据分布与全局数据分布的相似性,并基于相似性计算各客户端优先级分数,选择优先级较高的客户端进行通信。该策略可有效减少客户端与中心服务器的累计通信次数,提高通信效率并加速模型收敛。最后,仿真实验结果表明,相较于其他联邦学习算法,所提算法能够在保证高准确率的前提下,将累计通信次数减少至少50%。展开更多
Lung cancer is one of the malignant tumor diseases with high morbidity and high mortality in the world. Non-small cell lung cancer (NSCLC) is the most common pathological type of lung cancer. Currently, chemotherapy, ...Lung cancer is one of the malignant tumor diseases with high morbidity and high mortality in the world. Non-small cell lung cancer (NSCLC) is the most common pathological type of lung cancer. Currently, chemotherapy, targeted therapy, immunotherapy or combination therapy is the main treatment for NSCLC, but it is still inevitably faced with the challenges of acquired drug resistance and tumor progression. The birth of antibody conjugator provides a new choice for its treatment. Antibody conjugator is a new type of biotherapeutic drug which is connected by monoclonal antibody via linker and cytotoxic drug. It has the characteristics of precision, high efficiency and low toxicity, etc. In recent years, its research and development and clinical trials have been endless. It shows that this new type of drug has great potential in the field of tumor therapy. In this paper, the structural characteristics, mechanism of action, current application, research achievements, challenges, countermeasures and development of ADC in NSCLC treatment are reviewed.展开更多
文摘In non-independent and identically distributed(non-IID)data environments,model performance often degrades significantly.To address this issue,two improvement methods are proposed:FedReg and FedReg^(*).FedReg is a method based on hybrid regularization aimed at enhancing federated learning in non-IID scenarios.It introduces hybrid regularization to replace traditional L2 regularization,combining the advantages of L1 and L2 regularization to enable feature selection while preventing overfitting.This method better adapts to the diverse data distributions of different clients,improving the overall model performance.FedReg^(*)combines hybrid regularization with weighted model aggregation.In addition to the benefits of hybrid regularization,FedReg^(*)applies a weighted averaging method in the model aggregation process,calculating weights based on the cosine similarity between each client gradient and the global gradient to more reasonably distribute client contributions.By considering variations in data quality and quantity among clients,FedReg^(*)highlights the importance of key clients and enhances the model’s generalization performance.These improvement methods enhance model accuracy and communication efficiency.
文摘联邦学习因具有隐私保护的天然特性,已经逐渐成为一个被广泛认可的分布式机器学习框架。但由于参与方数据分布的差异性,特别是呈现非独立同分布(Non-Independent and Identically Distributed,Non-IID)时,其面临着泛化性能不足、收敛性能下降、数据倾斜等严峻挑战。用预训练基础模型缓解Non-IID问题作为一种新颖的方法,演变出了各种各样的解决方案。对此,从预训练基础模型的角度,对现有工作进行了综述。首先介绍了基础模型方法,对典型的基础模型编码结构进行对比分析。其次从修改输入、基础模型部分结构再训练,以及参数高效微调3个角度,提出了一种新的分类方法。最后探讨了该类工作的核心难题和未来研究方向。
文摘目的评估血清非高密度脂蛋白胆固醇与高密度脂蛋白胆固醇比值(non-high density lipoprotein cholesterol/high density lipoprotein cholesterol,NHHR)、血浆致动脉粥样硬化指数(atherogenic index of plasma,AIP)与慢性肾病(chronic kidney disease,CKD)发病风险的关系,为CKD的防治提供依据。方法以金昌队列中25377名未患CKD的参与者作为研究对象,采用Cox比例风险回归模型、限制性立方样条分别评估NHHR和AIP对CKD的发病风险及剂量-反应关系,并进行亚组分析。采用受试者工作特征曲线评估NHHR和AIP对CKD发病风险的预测能力。结果经过平均4.77年的随访调查后,新发CKD患者有1213例,发病密度为10.03/1000人年。调整混杂因素后,相较于Q1组,Q4组人群中NHHR和AIP的CKD发病风险比分别为1.270(95%CI:1.066~1.512)和1.294(95%CI:1.081~1.548),且均存在一定的剂量-反应关系(均P<0.05)。NHHR和AIP预测CKD的AUC值分别为0.750(95%CI:0.736~0.764)和0.735(95%CI:0.721~0.749)。亚组分析发现,吸烟和糖尿病与NHHR、糖尿病和AIP间存在交互作用(均P<0.05)。结论NHHR和AIP是CKD发病的独立危险因素,并对CKD发病风险有一定的预测能力。
文摘在客户端数据非独立同分布(Non‑Independent and Identically Distributed,Non‑IID)的场景中,为了向客户端提供个性化且通信高效的解决方案,提出了一种面向Non‑IID场景的通信高效个性化联邦学习算法。具体地,为充分利用相似客户端之间的知识提升模型性能,同时保留本地客户端的个性化信息,提出一种融合模型分层思想与聚类思想的个性化联邦学习算法。为进一步解决通信开销高的问题,设计选择性模型聚合策略,在中心服务器通过最大均值差异评估客户端数据分布与全局数据分布的相似性,并基于相似性计算各客户端优先级分数,选择优先级较高的客户端进行通信。该策略可有效减少客户端与中心服务器的累计通信次数,提高通信效率并加速模型收敛。最后,仿真实验结果表明,相较于其他联邦学习算法,所提算法能够在保证高准确率的前提下,将累计通信次数减少至少50%。
文摘Lung cancer is one of the malignant tumor diseases with high morbidity and high mortality in the world. Non-small cell lung cancer (NSCLC) is the most common pathological type of lung cancer. Currently, chemotherapy, targeted therapy, immunotherapy or combination therapy is the main treatment for NSCLC, but it is still inevitably faced with the challenges of acquired drug resistance and tumor progression. The birth of antibody conjugator provides a new choice for its treatment. Antibody conjugator is a new type of biotherapeutic drug which is connected by monoclonal antibody via linker and cytotoxic drug. It has the characteristics of precision, high efficiency and low toxicity, etc. In recent years, its research and development and clinical trials have been endless. It shows that this new type of drug has great potential in the field of tumor therapy. In this paper, the structural characteristics, mechanism of action, current application, research achievements, challenges, countermeasures and development of ADC in NSCLC treatment are reviewed.
