ML Evidence Recommendation for Dispute Resolution
Solution Overview
Problem
Current systems for dispute resolution in online transaction processing lack the ability to effectively recommend and facilitate the submission of relevant evidence, leading to inefficiencies and increased time and resources for both service providers and merchants in resolving disputes.
Innovation Solution
A machine learning system is developed to classify and recommend evidence types based on their likelihood of resolving disputes, utilizing trained ML models that classify evidence into categories and predict the probability of success, allowing merchants to submit the most effective evidence for dispute resolution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If current dispute resolution systems are used without evidence recommendation, then merchants must manually locate and upload evidence, but this causes unnecessary time and computing resources to be wasted on multiple evidence submission cycles
Solution Approach 1:
The system performs preliminary classification and recommendation of evidence types before the merchant actually submits evidence. The ML model analyzes the dispute description and pre-recommends appropriate evidence categories, allowing the merchant to prepare and upload the correct evidence in a single submission cycle rather than through multiple iterative attempts.
Solution Approach 2:
The patent introduces an intermediary evidence recommendation system that acts as a mediator between the dispute description and the evidence submission process. This intermediary ML-based system analyzes the dispute and bridges the gap by suggesting specific evidence types, reducing the manual effort required while maintaining ease of operation for the merchant.
2Productivity
If evidence recommendation system is implemented, then evidence submission efficiency is improved, but system complexity increases due to ML model training and deployment
Solution Approach 1:
The patent segments the evidence recommendation functionality into a separate, modular ML-based component that can be independently trained, deployed, and maintained. This segmentation allows the core dispute resolution system to remain relatively simple while the ML module handles the complex analysis and recommendation tasks, making the overall system complexity manageable through clear separation of concerns.
Solution Approach 2:
The ML model is trained on historical dispute and evidence data to autonomously learn and improve evidence recommendation capabilities without requiring manual configuration or intervention for each new dispute type. The system serves itself by continuously learning from past cases, reducing the operational complexity of managing the recommendation system while maintaining high productivity.
3Loss of energy
If merchants manually determine and locate evidence without recommendations, then system resources are conserved, but unnecessary computing resources are consumed from multiple evidence resubmissions
Solution Approach 1:
The system implements a feedback mechanism where the ML model learns from historical dispute resolution outcomes and evidence effectiveness data. This feedback loop allows the system to improve its recommendations over time, providing increasingly accurate information about proper evidence types to merchants, thereby reducing wasted computing resources on incorrect evidence submissions while continuously improving the quality of information provided.
Data Source
AI summary
There are provided systems and methods for a machine learning system for automated recommendations of evidence during dispute resolution. A service provider, such as an electronic transaction processor for digital transactions, may provide a dispute resolution system, which allows adjudication of disputes between merchants and customers. The dispute resolution system may employ a machine learning engine that performs evidence classification and recommendation using one or more machine learning models. A first model may be trained to classify evidence based on text and evidence categories. Using the classified evidence, a second model may be trained to recommend evidence that has a highest probability of winning a dispute from past dispute resolutions and those evidence categories submitted to the dispute. The second model may provide one or more evidence categories for a dispute party to submit via a user interface and may rank or otherwise suggest evidence by the user interface.