Prior studies commonly use an auditor's market share in an industry as a proxy for auditor industry expertise and find that audit quality is positively related to an audit partner's within-industry market shar...Prior studies commonly use an auditor's market share in an industry as a proxy for auditor industry expertise and find that audit quality is positively related to an audit partner's within-industry market share in a voluntary audit partner rotation regime where the length of the client-partner relationship is not limited.Mandatory audit partner rotation,however,limits the length of the client-partner relationship and can artificially increase or decrease the market shares of incoming and departing partners,thus making the audit partner's within-industry market share an unreliable proxy for auditor industry expertise.Using a sample of banks in Taiwan,China,we find that audit quality is positively related to an audit partner's within-industry market share in the voluntary audit partner rotation regime.However,such a positive relation disappears in the mandatory audit partner rotation regime.Thus,we conclude that mandatory audit partner rotation decouples the link between an audit partner's within-industry market share and auditor industry expertise and caution researchers against using an audit partner's within-industry market share as a proxy for auditor industry expertise in a mandatory audit partner rotation regime.展开更多
Federated learning(FL)has shown great potential in addressing data privacy issues in medical image analysis.However,varying data distributions across different sites can create challenges in aggregating client models ...Federated learning(FL)has shown great potential in addressing data privacy issues in medical image analysis.However,varying data distributions across different sites can create challenges in aggregating client models and achieving good global model performance.In this study,we propose a novel personalized contrastive representation FL framework,named PCRFed,which leverages contrastive representation learning to address the non-independent and identically distributed(non-IID)challenge and dynamically adjusts the distance between local clients and the global model to improve each client’s performance without incurring additional communication costs.The proposed weighted model-contrastive loss provides additional regularization for local models,optimizing their respective distributions while effectively utilizing information from all clients to mitigate performance challenges caused by insufficient local data.The PCRFed approach was evaluated on two non-IID medical image segmentation datasets,and the results show that it outperforms several state-of-the-art FL frameworks,achieving higher single-client performance while ensuring privacy preservation and minimal communication costs.Our PCRFed framework can be adapted to various encoder-decoder segmentation network architectures and holds significant potential for advancing the use of FL in real-world medical applications.Based on a multi-center dataset,our framework demonstrates superior overall performance and higher single-client performance,achieving a 2.63%increase in the average Dice score for prostate segmentation.展开更多
文摘Prior studies commonly use an auditor's market share in an industry as a proxy for auditor industry expertise and find that audit quality is positively related to an audit partner's within-industry market share in a voluntary audit partner rotation regime where the length of the client-partner relationship is not limited.Mandatory audit partner rotation,however,limits the length of the client-partner relationship and can artificially increase or decrease the market shares of incoming and departing partners,thus making the audit partner's within-industry market share an unreliable proxy for auditor industry expertise.Using a sample of banks in Taiwan,China,we find that audit quality is positively related to an audit partner's within-industry market share in the voluntary audit partner rotation regime.However,such a positive relation disappears in the mandatory audit partner rotation regime.Thus,we conclude that mandatory audit partner rotation decouples the link between an audit partner's within-industry market share and auditor industry expertise and caution researchers against using an audit partner's within-industry market share as a proxy for auditor industry expertise in a mandatory audit partner rotation regime.
基金supported by the National Key R&D Program of China,No.2023YFC2415200National Natural Science Foundation of China,Nos.82361168664,U24A20759,82441018,82372053,92259302,82302296+4 种基金Beijing Natural Science Foundation,Nos.JQ24048,Z20J00105Scientific and Technological Innovation Project of China Academy of Chinese Medical Sciences,No.CI2023C008YGScience and Technology Development Fund of Macao Special Administrative Region,No.0006/2023/AFJKey-Area Research and Development Program of Guangdong Province,No.2021B0101420005the Youth Innovation Promotion Association CAS,No.Y2021049.
文摘Federated learning(FL)has shown great potential in addressing data privacy issues in medical image analysis.However,varying data distributions across different sites can create challenges in aggregating client models and achieving good global model performance.In this study,we propose a novel personalized contrastive representation FL framework,named PCRFed,which leverages contrastive representation learning to address the non-independent and identically distributed(non-IID)challenge and dynamically adjusts the distance between local clients and the global model to improve each client’s performance without incurring additional communication costs.The proposed weighted model-contrastive loss provides additional regularization for local models,optimizing their respective distributions while effectively utilizing information from all clients to mitigate performance challenges caused by insufficient local data.The PCRFed approach was evaluated on two non-IID medical image segmentation datasets,and the results show that it outperforms several state-of-the-art FL frameworks,achieving higher single-client performance while ensuring privacy preservation and minimal communication costs.Our PCRFed framework can be adapted to various encoder-decoder segmentation network architectures and holds significant potential for advancing the use of FL in real-world medical applications.Based on a multi-center dataset,our framework demonstrates superior overall performance and higher single-client performance,achieving a 2.63%increase in the average Dice score for prostate segmentation.
