Recursive Decision Engines with Validated Real-Time Model Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a need for methods to deploy a recursive decisioning system that utilizes machine learning models and artificial intelligence, allowing for dynamic and automated updates in real-time as users interact with the system, while ensuring compliance with predefined guidelines and performance metrics.
Innovation Solution
A computer-implemented method for generating and deploying a machine learning-enabled decision engine (MLDE) that includes defining decision options, data inputs, and performance metrics, applying controls, and validating the MLDE, with the ability to retrain predictive models using updated datasets and generate synthetic data for continuous improvement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If machine learning models are continuously updated in real-time based on user interactions, then the adaptability and performance of the decision engine improve, but the system complexity and computational resources required increase
Solution Approach 1:
The system is divided into distinct modules: data collection component, model training component, validation component, and deployment component. Each module handles specific tasks independently, allowing the complex continuous update process to be managed through modular operations that can be executed separately and combined systematically.
Solution Approach 2:
The system performs preliminary validation of updated models before deploying them to production. A validation environment is used to test model performance and compliance with controls before the models are activated, preventing potentially harmful updates from reaching the live system and reducing the risk associated with continuous updates.
2Measurement precision
If machine learning models are continuously retrained with updated datasets, then the decision-making accuracy improves, but the time and computational resources required increase
Solution Approach 1:
The system implements periodic retraining cycles where models are retrained at scheduled intervals or when specific triggers are met (such as accumulating a certain amount of new data). This periodic approach balances the need for updated accuracy with the practical constraints of time and computational resources, rather than continuously retraining without interruption.
Solution Approach 2:
The system maintains continuous operation by validating and deploying model updates in a streamlined manner. The validation process is designed to quickly assess model performance, and once validated, updates are deployed to maintain continuous improved decision-making capability without significant interruptions to system operation.
3Reliability
If controls are applied to limit decision-making options, then compliance with guidelines is ensured, but the versatility of the decision engine is reduced
Solution Approach 1:
The controls applied to the decision engine are dynamic rather than static. The system can adjust the stringency and type of controls based on the specific context, data characteristics, and compliance requirements. This allows the system to maintain versatility by applying appropriate controls only when necessary, rather than imposing rigid constraints that would limit all decision options regardless of situation.
Data Source
AI summary
The present disclosure relates generally to the generation and deployment of a machine learning-enabled decision engine (MLDE). The MLDE includes decision options that are composed of a discrete list of selectable options. Further, the MLDE includes data inputs that can be used to influence decisions made by the machine learning models of the MLDE. Controls are applied to the MLDE to overlay and bound the decisioning within guidelines established by an operator of the MLDE. Once the MLDE is established, the MLDE is validated and deployed for use by software applications to make decisions.