MACHINE LEARNING SYSTEM FOR TAKING CONTROL ACTIONS
First Claim
1. A method of training machine learning models, in a data processing system comprising at least one processor and at least one memory, the at least one memory comprising instructions executed by the at least one processor to process transactions, the method comprising:
- receiving a transaction;
forwarding the transaction to at least one of a plurality of integrated control action models that use outputs of one model as inputs to other models, wherein the plurality of integrated control action models are machine learning models jointly trained for taking each control action of a plurality of control actions on the transaction to maximize an objective function based on a probability of the plurality of control actions matching corresponding target control actions taken on the transaction, wherein the plurality of integrated control action models include at least a risk model configured to output risk prediction information for a first control action that indicates whether or not to initiate processing of the transaction;
receiving the risk prediction information from the risk model; and
executing at least the first control action based on the risk prediction information.
1 Assignment
0 Petitions
Accused Products
Abstract
A device in a data processing system for training machine learning models receives a transaction and forwards it to at least one of a plurality of integrated control action models that use outputs of one model as inputs to other models. The models are machine learning models jointly trained for taking each control action of a plurality of control actions on the transaction to maximize an objective function based on probabilities of the control actions matching corresponding target control actions. The machine learning models include a risk model that outputs risk prediction information for a first control action that indicates whether or not to initiate processing of the transaction. The device further receives the risk prediction information from the risk model, and executes at least the first control action based on the risk prediction information.
29 Citations
21 Claims
-
1. A method of training machine learning models, in a data processing system comprising at least one processor and at least one memory, the at least one memory comprising instructions executed by the at least one processor to process transactions, the method comprising:
-
receiving a transaction; forwarding the transaction to at least one of a plurality of integrated control action models that use outputs of one model as inputs to other models, wherein the plurality of integrated control action models are machine learning models jointly trained for taking each control action of a plurality of control actions on the transaction to maximize an objective function based on a probability of the plurality of control actions matching corresponding target control actions taken on the transaction, wherein the plurality of integrated control action models include at least a risk model configured to output risk prediction information for a first control action that indicates whether or not to initiate processing of the transaction; receiving the risk prediction information from the risk model; and executing at least the first control action based on the risk prediction information. - View Dependent Claims (2, 3, 4, 5)
-
-
6. A method, in a data processing system comprising at least one processor and at least one memory, the at least one memory comprising instructions executed by the at least one processor to process transactions using machine learning, the method comprising:
-
receiving a current transaction; and processing the current transaction by taking at least one control action using a joint learned model, wherein the joint learned model comprises at least a first learned model and a second learned model that are jointly trained for taking each control action of a plurality of control actions on at least one training transaction so as to maximize an expected value of an objective function associated with the at least one training transaction based on a probability of the plurality of control actions matching corresponding target control actions taken on the at least one training transaction using at least the first learned model and the second learned model. - View Dependent Claims (7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19)
-
-
20. A device in a data processing system for training machine learning models, comprising:
-
at least one processor; and at least one memory in communication with the at least one processor, wherein the at least one memory comprises instructions executed by the at least one processor to process transactions, comprising; receiving a transaction; forwarding the transaction to at least one of a plurality of integrated control action models that use outputs of one model as inputs to other models, wherein the plurality of integrated control action models are machine learning models jointly trained for taking each control action of a plurality of control actions on the transaction to maximize an objective function based on a probability of the plurality of control actions matching corresponding target control actions taken on the transaction, wherein the plurality of integrated control action models include at least a risk model configured to output risk prediction information for a first control action that indicates whether or not to initiate processing of the transaction; receiving the risk prediction information from the risk model; and executing at least the first control action based on the risk prediction information.
-
-
21. A transaction processing apparatus, comprising:
-
a memory; and at least one processor coupled to the memory, wherein the at least one processor is configured to process transactions using machine learning by; receiving a current transaction; and processing the current transaction by taking at least one control action using a joint learned model, wherein the joint learned model comprises at least a first learned model and a second learned model that are jointly trained for taking each control action of a plurality of control actions on at least one training transaction so as to maximize an expected value of an objective function associated with the at least one training transaction based on a probability of the plurality of control actions matching corresponding target control actions taken on the at least one training transaction using at least the first learned model and the second learned model.
-
Specification