Reinforcement Learning Agent for Synthetic Financial Transaction Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Financial crime detection systems require large amounts of realistic simulated customer transaction data to improve predictive models, but sensitive real customer data is limited, making it challenging to effectively simulate fraudulent situations.
Innovation Solution
A computer-implemented method using a reinforcement learning model with an intelligent agent, policy engine, and environment to simulate transaction data by iteratively adjusting policies based on feedback, ensuring the similarity of simulated data to standard customer transaction data exceeds predefined thresholds, incorporating unsupervised clustering to acquire standard customer transaction data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If real customer data is used to train predictive models, then model accuracy is improved, but customer privacy is compromised and data availability is limited
Solution Approach 1:
The patent creates synthetic copies of customer transaction data that preserve the statistical properties and behavioral patterns of real data while containing no actual personal information. The simulation engine generates artificial transaction records that replicate fraudulent and legitimate transaction characteristics, enabling model training without exposing real customer data.
Solution Approach 2:
The patent introduces a simulation engine as an intermediary between real customer data and the predictive model training process. This intermediary component processes and transforms real data patterns into synthetic data, acting as a buffer that prevents direct use of sensitive information while maintaining data utility for model development.
2Productivity
If more simulated customer data is generated to improve model training, then predictive model performance is improved, but data realism and authenticity may deteriorate
Solution Approach 1:
The patent implements a feedback mechanism where the simulation engine continuously compares generated synthetic data against real transaction data patterns and adjusts its generation process accordingly. This feedback loop ensures that the simulated data maintains statistical fidelity and behavioral authenticity, preventing degradation of data realism even as large volumes of synthetic data are produced.
Solution Approach 2:
The patent employs parameter-based control of the simulation engine to adjust the characteristics of generated data. By modifying parameters such as transaction frequency, amount distributions, and behavioral patterns, the system can generate diverse yet realistic synthetic data that accurately reflects various customer types and transaction scenarios while maintaining statistical validity.
Data Source
AI summary
Embodiments can provide a computer implemented method for simulating transaction data using a reinforcement learning model including an intelligent agent, a policy engine, and an environment, the method including: providing standard customer transaction data representing a group of customers having similar transaction characteristics as a goal; and performing a plurality of iterations to simulate the standard customer transaction data, wherein the plurality of iterations is performed until a degree of similarity of simulated customer transaction data relative to the standard customer transaction data is higher than a first predefined threshold. In each iteration, the method includes: conducting, by the intelligent agent, an action including a plurality of simulated transactions; comparing, by the environment, the action with the goal; providing by the environment, a feedback associated with the action based on a degree of similarity relative to the goal; and adjusting, by the policy engine, a policy based on the feedback.