Synthetic Data Detection System for Fraud Privacy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing detection systems for telephony fraud and service attacks face challenges in building general classification models due to the need to choose between a wide range of customer data and a wide feature space, often resulting in subpar performance due to the inclusion of proprietary and personally-identifiable information.
Innovation Solution
The use of synthetic communications session data generated by neural networks, trained using proprietary customer data, to create a detection system that determines the probability of malicious transactions without revealing actual customer data, allowing for the implementation of mitigation actions based on the determined probability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If actual customer data is used to train detection systems, then detection performance is improved, but customer privacy and data security are compromised
Solution Approach 1:
The patent uses Generative Adversarial Networks (GANs) to create synthetic customer data that copies the statistical properties and patterns of actual customer data without containing any real customer information. The synthetic data preserves the feature distributions and relationships needed for training detection systems while eliminating privacy risks associated with using real customer data.
Solution Approach 2:
The patent introduces synthetic data as an intermediary between actual customer data and the detection system training process. The GAN-based synthetic data generation acts as a mediator that transfers the essential characteristics of real data to the training set without exposing actual customer information, thus resolving the contradiction between detection performance and privacy protection.
2Adaptability or versatility
If data from multiple customers is aggregated for training, then generalization capability is improved, but data heterogeneity and quality control become more difficult
Solution Approach 1:
The patent transforms the data aggregation problem by changing the parameter space from raw customer data to synthetic data parameters. By generating synthetic data that matches the statistical parameters (distributions, correlations, feature relationships) of multiple customers, the system achieves generalization capability without the complexity of directly aggregating and harmonizing heterogeneous real data from multiple sources.
Data Source
AI summary
Methods and apparatus for implementing and operating malicious transaction detection systems. An exemplary method embodiment includes the steps of: (i) operating, a malicious transaction detection system, to receive communications session establishment data; operating, the malicious transaction detection system, to determine a probability of whether or not the communications session establishment data indicates that the communications session is malicious; and when the determined probability is greater than or equal to a predetermined threshold value determining that a transaction corresponding to the received communications session establishment data is malicious; and when the determined probability is less than the predetermined threshold value determining that the transaction corresponding to the received communications session establishment data is not malicious; and wherein the malicious transaction detection system includes a determination model trained using synthetic communications session data.