💣The Open Decision Intelligence Platform (DIP) should allow all levels of intelligence in organizations based on domain knowledge of the sector.
💢Data-driven intelligence #machinelearning #dealpearning #llms based on transformer architecture cannot achieve reliable intelligence that is explainable, trustworthy, compliant, and unbiased while it does valid reasoning.
The only way to achieve that is to use neuro-symbolic #ai, which addresses the weakness of data-driven while achieving cognitive intelligence with significant reasoning based on domain knowledge.
Decision Intelligence Architecture provides a roadmap for the layers of enterprise intelligence 👇
– Knowledge Core
– Intelligence Core
– Expert Intelligence
– Cognitive Intelligence
– Autonomous Intelligence
✨Each layer is a prerequisite to the previous layer. More importantly, the Intelligence Core provides 100% explainability, transparency, and accuracy for a regulated and dynamic environment where data may not even exist to train a #machinelearning model.
In addition, the “Intelligence Core” should be based on the Decision Model and Notation (DMN), which provides interoperability across platforms. To do that, the ” Intelligence Core” should become Conformance Level 3 (CL3) with Decision Model and Notation, which allows fully executable decision-making models based on Symbolic #AI.
🚀If you think about it, this sounds magical, but it essentially is like intelligence reusability: sharing knowledge-driven intelligence between systems and processes based on an open standard!
#bfsi #insruance #financialservices #banking #government #decisionintelligence
Posted here.
Published May 31st, 2024 at 07:30 am
