Heterogeneous ML Model Integration in RPA Workflows
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current robotic process automation (RPA) and machine learning (ML) technologies face barriers in seamless integration due to operational, technological, and process disconnects, leading to challenges in validating and deploying ML models within RPA workflows, resulting in a lack of transparency and effective utilization of ML models.
Innovation Solution
A computer program and method for validating ML models, uploading them for storage, and deploying them via a REST API, enabling seamless integration and management of ML models within RPA workflows through a drag-and-drop interface, model versioning, and CRUD operations, facilitating their consumption within RPA workflows.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If ML models are published via external platforms (Azure, Google, Amazon) over HTTP API, then ML model accessibility is improved, but integration complexity with RPA workflows increases due to disconnected tools and pipelines
Solution Approach 1:
The patent merges ML model publishing and RPA workflow execution into a single integrated platform. The RPA platform now natively supports ML model deployment, validation, and consumption without requiring external platforms, thereby reducing integration complexity while maintaining model accessibility.
Solution Approach 2:
The patent introduces an intermediary validation mechanism that sits between ML model publishing and RPA workflow consumption. This validation layer ensures model correctness and compatibility before integration, reducing the complexity of direct integration with external platforms.
2Stability of the object's composition
If RPA and ML are managed as separate processes with independent teams, then operational independence is maintained, but transparency and model validation capability deteriorate
Solution Approach 1:
The patent combines RPA and ML management into a unified platform while maintaining distinct functional modules. This allows independent teams to work on their respective components while achieving transparency through shared validation mechanisms and common deployment pipelines.
Solution Approach 2:
The patent implements feedback loops where ML models are automatically validated within the RPA workflow context. The validation results provide transparency about model correctness and compatibility, allowing both RPA and ML teams to see the impact of their work on the other.
3Adaptability or versatility
If code writing is required to consume ML APIs, then flexibility in model consumption is improved, but ease of operation for RPA developers deteriorates
Solution Approach 1:
The patent provides pre-built templates and standardized interfaces for consuming ML models within RPA workflows. Instead of requiring developers to write custom code for each API consumption, they can use validated, pre-configured connection patterns that maintain flexibility while significantly improving ease of operation.
4Productivity
If ML models are deployed without validation, then deployment speed is improved, but reliability of model output deteriorates
Solution Approach 1:
The patent performs ML model validation as a preliminary step before deployment to RPA workflows. The validation process checks model correctness, compatibility, and performance metrics in advance, ensuring reliable model output while maintaining fast deployment through automated validation routines.
Data Source
AI summary
Frameworks and techniques for integration of heterogeneous machine learning (ML) models into robotic process automation (RPA) workflows are provided. This may be accomplished via a seamless drag-and-drop interface that allows deployment of ML models into an RPA workflow. Via a framework, these heterogeneous models may be provided by customers, third parties, and/or partners and integrated into the RPA workflow. The framework may provide a straightforward way to deploy machine learning models via a conductor and to manage model versioning and create/retrieve/update/delete (CRUD). The framework may facilitate integration of different models into the RPA workflow through the steps of uploading, validating, publishing, and deploying models.