ML Bill Splitting Engine for Automated Transaction Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The process of splitting bills in social or group transactions remains largely manual, inconvenient, and often requires manual sorting and calculation, especially in scenarios where electronic payment systems do not facilitate easy distribution among participants.
Innovation Solution
A machine learning architecture that processes raw input data, such as images of bills or social media posts, to identify participants and calculate individual contributions, using OCR, facial recognition, and historical transaction data to recommend and facilitate electronic bill splitting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual bill splitting is used, then accuracy in calculating individual contributions can be maintained, but time consumption and convenience deteriorate
Solution Approach 1:
The patent replaces manual mechanical calculation and sorting processes with an automated electronic system that uses image recognition, OCR, and machine learning algorithms to automatically process bills, identify participants, and calculate contributions, eliminating the need for manual intervention
Solution Approach 2:
The system enables automatic self-service bill splitting by capturing transaction data through images or electronic receipts, automatically identifying participants via facial recognition or contact information, and automatically distributing payment amounts without requiring manual calculation or coordination among participants
2Speed
If electronic payment systems are used, then speed of transaction can be improved, but ability to handle complex bill splitting scenarios deteriorates
Solution Approach 1:
The system dynamically adjusts processing parameters based on the complexity of the bill splitting scenario, using machine learning models to adapt to different transaction types, participant configurations, and payment distributions, enabling flexible handling of diverse scenarios while maintaining speed
Solution Approach 2:
The patent segments the bill splitting process into distinct automated stages: image capture and processing, participant identification through facial recognition or contact matching, calculation of individual contributions based on itemized charges, and electronic payment distribution, allowing each stage to be optimized independently for speed and accuracy
3Productivity
If automated bill splitting is implemented, then convenience and speed are improved, but system complexity increases
Solution Approach 1:
The patent implements a universal bill splitting platform that handles multiple transaction types (dining, shopping, entertainment), various participant configurations, and different payment methods through a single integrated system, reducing the need for multiple specialized systems and simplifying overall complexity
Data Source
AI summary
Disclosed herein are system, method, and computer program product embodiments for providing recommendations for splitting bills. The approaches disclosed include the ability to obtain information about a bill to be split (such as a photo of the bill), and then use several machine learning models to determine the ‘who,’‘what,’ and ‘where’ of the underlying transaction. In particular, machine learning models described herein are used to perform facial recognition of a ‘selfie’ taken when a transaction was made against social media accounts to determine participants of the transaction. The machine learning models may also identify expected pricing from data about a merchant associated with the transaction, and expected amounts for each participant based on the expected pricing.