Embedded Machine Learning for Enterprise Process Integration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional machine learning models are cryptic and require expertise to consume, leading to integration issues with business processes, compliance challenges, and difficulties in updating or managing them within enterprise systems, particularly in ensuring GDPR compliance and security.
Innovation Solution
An embedded machine learning architecture that integrates machine learning models directly with entity business processes on computing systems, allowing all users to access and manage them, leveraging in-memory database platforms like SAP HANA for enhanced speed and compliance, and utilizing predictive analytics libraries for low-resource algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If machine learning models are stored in high-end computing systems separately from business processes, then model performance and computational power are improved, but integration with business processes deteriorates and compliance security worsens
Solution Approach 1:
The patent merges machine learning models directly into business process software packages, allowing the models to be stored and executed within the same system environment as the business processes they support. This integration enables seamless embedding of ML functionality into enterprise resource planning systems without requiring separate high-end computing infrastructure, thus maintaining both computational capability and process integration.
2Measurement precision
If machine learning models are stored in high-end computing systems, then model accuracy is improved, but accessibility for non-expert users deteriorates and management complexity increases
Solution Approach 1:
The patent introduces an intermediary layer that packages machine learning models within standardized software containers that can be easily deployed and managed through conventional software distribution mechanisms. This intermediary packaging approach allows non-expert users to access and utilize ML models through familiar software interfaces without needing to understand the underlying model complexity or require specialized expertise.
3Reliability
If machine learning models are updated externally, then model performance is improved, but compliance security and data protection worsen
Solution Approach 1:
The patent implements preliminary action by embedding compliance and security controls directly into the model packaging and deployment process. Rather than addressing compliance after models are updated externally, the system pre-configures security measures, data protection mechanisms, and compliance validations as integral parts of the model software package before deployment, ensuring that performance updates do not compromise security or regulatory requirements.
4Ease of operation
If machine learning models are integrated into business processes, then ease of management is improved, but system resource requirements worsen
Solution Approach 1:
The patent applies local quality by optimizing model deployment to use only the specific computational resources required for each individual model's operation within the business process environment. Rather than requiring uniform high-end infrastructure for all models, the system allows models to be deployed with resource allocations matched to their actual needs, enabling efficient use of local system resources while maintaining ease of management through standardized integration.
Data Source
AI summary
Systems and methods are provided for receiving a request for data associated with a particular functionality of an application, identifying a first attribute for which data is to be generated to fulfill the request, and determining that the first attribute corresponds to data to be generated by a first machine learning model. The systems and methods further providing for executing a view or procedure to generate data for input to the first machine learning model, inputting the generated data into the first machine learning model, and receiving output from the first machine learning model. The output is provided in response to the request for data associated with the particular functionality of the application.