Combining decisions with AI prediction
Statistical learning, especially in its latest deep learning (DL) form, has been trending more and more recently and offers advances in the general field of AI, which is very complementary to symbolic AI.
Machine learning (ML) is very good at perceiving—for example—image recognition or speech recognition, as well as at classifying unstructured data. ML models are also valuable for building predictors. With the assumption that the future is likely to behave like the past, we can build models to predict customer churn, borrower propensity to default on their loan payments, and other indicators derived from analysis of past data.
On the other hand, symbolic AI—such as the type of decision logic that can be expressed with rules or decision tables in Cloud Pak for Business Automation—is the preferred way to decide on and take explainable actions that can be verified for compliance with regulations or company policies...
