<div style="text-align:justify;"> In distributed AI system, the models trained on data from potentially unreliable sources can be attacked by manipulating the training data distribution by inserting ca...<div style="text-align:justify;"> In distributed AI system, the models trained on data from potentially unreliable sources can be attacked by manipulating the training data distribution by inserting carefully crafted samples into the training set, which is known as Data Poisoning. Poisoning will to change the model behavior and reduce model performance. This paper proposes an algorithm that gives an improvement of both efficiency and security for data poisoning in a distributed AI system. The past methods of active defense often have a large number of invalid checks, which slows down the operation efficiency of the whole system. While passive defense also has problems of missing data and slow detection of error source. The proposed algorithm establishes the suspect hypothesis level to test and extend the verification of data packets and estimates the risk of terminal data. It can enhance the health degree of a distributed AI system by preventing the occurrence of poisoning attack and ensuring the efficiency and safety of the system operation. </div>展开更多
The paper firstly gives a brief introduction of the credit focus as a result of the competition among the branch banks of the Commercial Bank. Then from the viewpoint of the performance of the branch bank, an analysis...The paper firstly gives a brief introduction of the credit focus as a result of the competition among the branch banks of the Commercial Bank. Then from the viewpoint of the performance of the branch bank, an analysis is made on the mechanism of the formation of the credit focus. Based on the analysis, the paper points out that the Commercial Bank branches, under competition of the deposit restrain, have brought about excessive credit, which exceeds the finance required by the normal business. And finally the paper forms the credit risk of the bank.展开更多
Finding successors for private enterprises has become an urgent problem in recent years, partially because those firms＇ owners lack trust in professional managers. Previous research on agents focuses on preventing op...Finding successors for private enterprises has become an urgent problem in recent years, partially because those firms＇ owners lack trust in professional managers. Previous research on agents focuses on preventing opportunistic behavior and neglects the value of their entrepreneurship. In our research, professional managers＇ credit is divided into three dimensions： personal credit, professional credit and operational credit. Using a sample of 379 firm owners from 27 provinces in China, we find that credit and its detailed dimensions are positively related to private owners＇ trust in professional managers. We also found that a rigorous and effective credit identification mechanism positively moderate the above relationships. Furthermore, trust affects professional managers＇ work performance in a positive way. Accordingly, we advise that （1） professional managers＇ credit assessment system should be established and improved; （2） professional managers and enterprise owners should attach importance to enhancing professionalism and promoting rigorous credit identification mechanisms; （3） the closed-loop of the credit-trust psychologically interactive mechanism based on credit mechanisms and credit identification mechanisms should be put to use.展开更多
文摘<div style="text-align:justify;"> In distributed AI system, the models trained on data from potentially unreliable sources can be attacked by manipulating the training data distribution by inserting carefully crafted samples into the training set, which is known as Data Poisoning. Poisoning will to change the model behavior and reduce model performance. This paper proposes an algorithm that gives an improvement of both efficiency and security for data poisoning in a distributed AI system. The past methods of active defense often have a large number of invalid checks, which slows down the operation efficiency of the whole system. While passive defense also has problems of missing data and slow detection of error source. The proposed algorithm establishes the suspect hypothesis level to test and extend the verification of data packets and estimates the risk of terminal data. It can enhance the health degree of a distributed AI system by preventing the occurrence of poisoning attack and ensuring the efficiency and safety of the system operation. </div>
文摘The paper firstly gives a brief introduction of the credit focus as a result of the competition among the branch banks of the Commercial Bank. Then from the viewpoint of the performance of the branch bank, an analysis is made on the mechanism of the formation of the credit focus. Based on the analysis, the paper points out that the Commercial Bank branches, under competition of the deposit restrain, have brought about excessive credit, which exceeds the finance required by the normal business. And finally the paper forms the credit risk of the bank.
文摘Finding successors for private enterprises has become an urgent problem in recent years, partially because those firms＇ owners lack trust in professional managers. Previous research on agents focuses on preventing opportunistic behavior and neglects the value of their entrepreneurship. In our research, professional managers＇ credit is divided into three dimensions： personal credit, professional credit and operational credit. Using a sample of 379 firm owners from 27 provinces in China, we find that credit and its detailed dimensions are positively related to private owners＇ trust in professional managers. We also found that a rigorous and effective credit identification mechanism positively moderate the above relationships. Furthermore, trust affects professional managers＇ work performance in a positive way. Accordingly, we advise that （1） professional managers＇ credit assessment system should be established and improved; （2） professional managers and enterprise owners should attach importance to enhancing professionalism and promoting rigorous credit identification mechanisms; （3） the closed-loop of the credit-trust psychologically interactive mechanism based on credit mechanisms and credit identification mechanisms should be put to use.
