Merchant-Specific Identifier Authentication via Machine Learning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

The existing payment processing systems relying on broad Merchant Category Codes (MCCs) lead to inefficiencies in reward allocations, security issues, and incorrect transaction authentication or denial, particularly due to the manual process of assigning correct MCCs to evolving businesses.

Innovation Solution

Implementing an automated and real-time processing system that uses Merchant Identifiers (MIDs) specific to each merchant, with a machine learning-based backend component for continuously updating lists of authenticated or prohibited merchants, enabling real-time transaction authentication or decline decisions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If broad Merchant Category Codes (MCCs) are used for transaction classification, then merchants can be grouped into categories for reward allocations and rule definitions, but this leads to incorrect authentication or denial of transactions and inefficiencies in reward allocations

Engineering Contradiction:

Improveability to group merchants into categoriesVSAvoidaccuracy of transaction authentication

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent segments the broad MCC classification system into more granular Merchant Specific Identifiers (MSIs). Each MSI uniquely identifies a specific merchant or a narrow group of merchants, replacing the coarse-grained MCC approach. This segmentation allows the system to maintain category-based functionality while achieving precise merchant identification, thereby resolving the contradiction between categorization capability and authentication accuracy.

Inventive Principle:

Principle #1Segmentation

2Ease of manufacture

If manual processes are used to assign MCCs to merchants, then MCC assignments can be made initially, but this becomes inefficient when businesses evolve and new businesses emerge

Engineering Contradiction:

Improvesimplicity of initial MCC assignmentVSAvoidefficiency of updating merchant classifications

Core Design Contradiction:

Ease of manufactureVSProductivity

Solution Approach 1:

The patent implements a self-updating system where Merchant Specific Identifiers are automatically generated and maintained through machine learning models that process transaction data. The system continuously learns and adapts to business evolutions without requiring manual intervention. This self-service mechanism replaces the inefficient manual MCC reassignment process, enabling the system to automatically adapt when businesses evolve or new businesses emerge.

Inventive Principle:

Principle #25Self-service

3Stability of the object's composition

If MCC codes are used for transaction processing, then existing legacy systems can be maintained, but this results in security issues and inconsistent denial or approval of transactions

Engineering Contradiction:

Improvecompatibility with legacy systemsVSAvoidsecurity vulnerabilities and inconsistent authentication

Core Design Contradiction:

Stability of the object's compositionVSObject-affected harmful factors

Solution Approach 1:

The patent introduces Merchant Specific Identifiers (MSIs) as an intermediary layer between the legacy MCC system and the transaction processing logic. The MSIs maintain compatibility with existing systems while enabling more precise and secure authentication decisions. This intermediary approach allows the system to leverage legacy infrastructure while eliminating the security vulnerabilities and inconsistencies associated with broad MCC-based authentication.

Inventive Principle:

Principle #24Intermediary (Mediator)

4Measurement precision

If real-time transaction authentication is implemented using machine learning models, then accurate authentication decisions can be made, but this increases processing complexity

Engineering Contradiction:

Improveaccuracy of authentication decisionsVSAvoidcomplexity of processing system

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent performs preliminary actions by pre-training machine learning models offline to generate Merchant Specific Identifiers and establish authentication rules before real-time transaction processing. The models are trained on historical transaction data to learn merchant patterns and behaviors in advance. During real-time processing, the system only needs to apply the pre-computed MSIs and trained model parameters, significantly reducing the computational complexity at transaction time while maintaining high authentication accuracy.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20230060331A1Automated authentication system based on target-specific identifier

Publication Date: 2023.03.02 SYNCHRONY BANK

AI summary

The disclosure is directed to a continuously and automatically updated authentication mechanism for authentication, in real-time, the processing of transactions at point of sales devices based on corresponding target-specific identifiers. In one aspect, a processing server includes one or more memories having computer-readable instructions stored therein, and one or more processors. The one or more processors are configured to execute the computer-readable instructions to receive a request for processing a transaction; identify a merchant-specific identifier for a merchant associated with the transaction; determine, in real-time and using a machine trained model, whether the merchant-specific identifier is a valid merchant-specific identifier or not; and process the transaction based on whether the machine trained model indicates that the merchant-specific identifier is valid or not.