Integration of Machine Learning and Artificial Intelligence with Decision Models
Historically decision models have been used to capture and automate business logic using decision tables and prescribed business rules. This application has made them an invaluable aid when the business logic is already well understood. However, integrating Artificial Intelligence (AI) or machine learning (ML) into your decision models can help if:
- Humans don’t know what the rules are or should be (e.g., predicting customer churn).
- Manual rule capture is impractical because the rules are very complex or highly volatile (e.g., making product recommendations based on prior purchasing behaviour which is constantly evolving).
- We need the decision model to learn from experience (e.g., selecting a web advertisement to display).
The combination of AI/ML with decision tables and business rules is compelling because:
- It allows the knowledge of human experts to be combined with statistical insights, provided by machine learning, promoting the stability, robustness and safety of AI.
- It provides a framework in which multiple AI models can be used together and their outputs combined to achieve a single goal.
- It supports the marshaling and control of AI models by human oversight.
- The framework of decision models improves the rigour and explainability of AI models.
- It encourages the development of AI and ML models to be closely driven by documented and measurable business goals.
RapidGen Genius allows your decision models to be integrated with existing or new AI or ML assets enabling you to define powerful decision models that synthesize AI, human expertise and robust oversight. The RapidGen architecture also makes models that are fully accountable for their outcomes.