Abstract: This specification discloses techniques for risk identification. One example method includes receiving, by a client device, a risk identification request identifying a requested service operation and service data associated with the requested service operation; retrieving, by the client device, service data corresponding to the risk identification request; determining, by the client device, service indicator data associated with the service data; analyzing, by the client device, one or more of the service data and the service indicator based on a risk identification rule or a risk identification model to produce a risk result; and determining, by the client device, whether the requested service operation is a high risk operation based at least in part on the risk result.

Abstract: This specification discloses techniques for risk identification. One example method includes receiving, by a client device, a risk identification request identifying a requested service operation and service data associated with the requested service operation; retrieving, by the client device, service data corresponding to the risk identification request; determining, by the client device, service indicator data associated with the service data; analyzing, by the client device, one or more of the service data and the service indicator based on a risk identification rule or a risk identification model to produce a risk result; and determining, by the client device, whether the requested service operation is a high risk operation based at least in part on the risk result.

Abstract: The present application discloses techniques for processing service requests based on risk. One example method includes: receiving a service processing request including service data; performing risk identification on the service processing request based on a stored risk identification rule or a stored risk identification model with reference to the service data to generate a risk identification result; determining that the risk identification result is an indefinite result; and in response to determining that the risk identification result is the indefinite result, sending a risk identification request that comprises the service data to a cloud risk identification system, wherein the risk identification request is configured to request the cloud risk identification system to perform risk identification on the service processing request.

Abstract: The present application discloses example methods, computer-readable mediums, and systems for controlling service operation risks. In one example, method, an application program on an end-user device monitors a service operation initiated by a user for invoking offline service information. After the service operation is monitored, it is determined whether the service operation is a risky operation based on recorded historical operation data and at least one of a predetermined risk evaluation rule or a risk evaluation model. In response to determining that the service operation is a risky operation, a refusal to invoke the offline service information is performed. In response to determining that the service operation is not a risky operation, the offline service information is invoked.

Abstract: This specification discloses techniques for risk identification. One example method includes receiving, by a client device, a risk identification request identifying a requested service operation and service data associated with the requested service operation; retrieving, by the client device, service data corresponding to the risk identification request; determining, by the client device, service indicator data associated with the service data; analyzing, by the client device, one or more of the service data and the service indicator based on a risk identification rule or a risk identification model to produce a risk result; and determining, by the client device, whether the requested service operation is a high risk operation based at least in part on the risk result.

Abstract: Implementations of the present specification disclose a method and an apparatus for detecting model security and an electronic device. The method includes the following: obtaining result data computed by using a model, wherein the result data comprises intermediate result data and output result data; obtaining a discriminator through training by using a generative adversarial network (GAN) framework, the model, and a sample; discriminating the result data by using the discriminator by using the model for current input data, to detect whether the model is currently secure, wherein the model is in a trusted execution environment for obtaining the discriminator through training; and determining a security detection result of the model.

Abstract: The present application discloses example methods, computer-readable mediums, and systems for controlling service operation risks. In one example, method, an application program on an end-user device monitors a service operation initiated by a user for invoking offline service information. After the service operation is monitored, it is determined whether the service operation is a risky operation based on recorded historical operation data and at least one of a predetermined risk evaluation rule or a risk evaluation model. In response to determining that the service operation is a risky operation, a refusal to invoke the offline service information is performed. In response to determining that the service operation is not a risky operation, the offline service information is invoked.

Abstract: Implementations of the present specification disclose a method and an apparatus for detecting model security and an electronic device. The method includes the following: obtaining result data computed by using a model, wherein the result data comprises intermediate result data and output result data; obtaining a discriminator through training by using a generative adversarial network (GAN) framework, the model, and a sample; discriminating the result data by using the discriminator by using the model for current input data, to detect whether the model is currently secure, wherein the model is in a trusted execution environment for obtaining the discriminator through training; and determining a security detection result of the model.

Abstract: A request message requesting a service is received from a terminal device. Computer code is transmitted to the terminal device, and a communication connection request is received from the terminal device in response to running the computer code on the terminal device. In response to receiving the communication connection request, a communication connection is established to the terminal device. A first IP address associated with the terminal device is determined based on one or more communications transmitted through the communication connection.

Abstract: The present application discloses example methods, computer-readable mediums, and systems for controlling service operation risks. In one example, method, an application program on an end-user device monitors a service operation initiated by a user for invoking offline service information. After the service operation is monitored, it is determined whether the service operation is a risky operation based on recorded historical operation data and at least one of a predetermined risk evaluation rule or a risk evaluation model. In response to determining that the service operation is a risky operation, a refusal to invoke the offline service information is performed. In response to determining that the service operation is not a risky operation, the offline service information is invoked.

Abstract: The present application discloses techniques for processing service requests based on risk. One example method includes: receiving a service processing request including service data; performing risk identification on the service processing request based on a stored risk identification rule or a stored risk identification model with reference to the service data to generate a risk identification result; determining that the risk identification result is an indefinite result; and in response to determining that the risk identification result is the indefinite result, sending a risk identification request that comprises the service data to a cloud risk identification system, wherein the risk identification request is configured to request the cloud risk identification system to perform risk identification on the service processing request.

Abstract: This specification discloses techniques for risk identification. One example method includes receiving, by a client device, a risk identification request identifying a requested service operation and service data associated with the requested service operation; retrieving, by the client device, service data corresponding to the risk identification request; determining, by the client device, service indicator data associated with the service data; analyzing, by the client device, one or more of the service data and the service indicator based on a risk identification rule or a risk identification model to produce a risk result; and determining, by the client device, whether the requested service operation is a high risk operation based at least in part on the risk result.

Abstract: A request message requesting a service is received from a terminal device. Computer code is transmitted to the terminal device, and a communication connection request is received from the terminal device in response to running the computer code on the terminal device. In response to receiving the communication connection request, a communication connection is established to the terminal device. A first IP address associated with the terminal device is determined based on one or more communications transmitted through the communication connection.

Abstract: Systems and methods for risk and fraud mitigation are presented. According to one or more aspects of the disclosure, a plurality of transactions may be processed without applying one or more active fraud rules. A hit rate for at least one fraud rule of the one or more active fraud rules then may be determined. Thereafter, a positive hit rate for the at least one fraud rule may be determined based on fraud event data corresponding to the plurality of transactions. In some arrangements, each transaction of the plurality of transactions may be randomly selected, from a larger plurality of received transactions, for inclusion in the plurality of transactions to be processed without application of the one or more active fraud rules. Additionally or alternatively, in certain arrangements, one or more fraud rules may be activated or deactivated based on their corresponding hit rates and positive hit rates.

Abstract: Systems and methods for risk and fraud mitigation are presented. According to one or more aspects of the disclosure, a plurality of transactions may be processed without applying one or more active fraud rules. A hit rate for at least one fraud rule of the one or more active fraud rules then may be determined. Thereafter, a positive hit rate for the at least one fraud rule may be determined based on fraud event data corresponding to the plurality of transactions. In some arrangements, each transaction of the plurality of transactions may be randomly selected, from a larger plurality of received transactions, for inclusion in the plurality of transactions to be processed without application of the one or more active fraud rules. Additionally or alternatively, in certain arrangements, one or more fraud rules may be activated or deactivated based on their corresponding hit rates and positive hit rates.