Abstract: Computer security improvements relating to detection of social engineering attacks using a deep learning pipeline for account and communication correlations are disclosed. A service provider may utilize a framework having computing operations for detecting social engineering attacks, fraud, and other malicious or suspicious activities by malicious bots and other fraudsters. In this regard, the service provider may extract pattern data from communications using one or more AI models, which may include semantic information. Account assets for different accounts may be correlated using a relationship graph that links the assets based on the communication patterns shared between the assets. The relationship graph may then be used by additional AI models that classify if the communications are associated with social engineering attacks by classifying the communications using an attack classifier. The classification may be based on inferencing by the AI models from training on past social engineering attacks.

Abstract: The present specification provides an outline binding method and apparatus, and a storage medium. The method includes: obtaining a query statement; generating a formatted query statement corresponding to the query statement; determining a target outline that matches the formatted query statement from at least one first-type outline, the first-type outline being a formatted outline to which a plurality of query statements capable of being normalized are jointly bound; and generating a plan corresponding to the query statement by using the target outline. The present specification allows a plurality of query statements to be jointly bound to a formatted outline, thereby implementing binding of a formatted outline without a need to modify service logic, improving flexibility of outline binding, and ensuring correctness and reliability of outline effectiveness.

Abstract: A method includes receiving, by a processor of a transaction processing entity, a transaction attempt. The method includes receiving a risk score from a risk strategy decision model, the risk score being determined from a machine learning model. The method includes in response to receiving the risk score, determining, whether the risk score exceeds a threshold indicating the transaction attempt is potentially fraudulent. In response to determining the risk score exceeds the threshold: the method includes determining, whether to approve or decline the transaction attempt; and determining a reason for approving or declining the transaction attempt based on one or more variables contributing to the risk score. The method includes outputting an indication to approve or decline the transaction attempt in response to determining whether to approve or decline the transaction attempt and the reason for approving or declining the transaction attempt.

Abstract: Various systems, mediums, and methods may perform operations, such as collecting various types of data from one or more data sources. Further, the operations may include learning user behaviors based on iterations of the collected historical data with a recurrent neural network (RNN) with long short term memory (LSTM). Yet further, the operations may include determining one or more feature vectors that represents the learned user behaviors. In addition, the operations may include generating one or more models associated with the learned user behaviors based on the one or more determined vectors.