Operationalizing machine learning ‘AIML’ is difficult
Dependencies in background knowledge create validation challenges:
These validation challenges must be overcome to operationalize AIML
without compromising system confidence, adaptability or robustness.
Organizations are investing heavily in enterprise BI systems
analyze and predict. These systems define organizational knowledge. AIML
appears to offer better predictive performance but is disconnected from
existing BI systems The challenge is why and how to
extend BI systems with AIML.
Organizations invest in machine learning data catalogs MLDCs
to better access enterprise data
The problem is MLDCs do not provide a formal basis
for labelling or attributing catalog entries:
Using the world’s first ontological computing language Blender Logic compares an organization’s own knowledge — as found in its enterprise BI systems. With the entities that comprise the inputs and outputs of its ongoing AIML processes. To provide validation services critical to the principled orchestration of AIML processes