How can ontologies serve as guardrails for AI Agents?

Ontologies provide a vocabulary (entities and relations) for your domain. The terms provide the basis for specifying constraints and rules that are the actual guardrails.

Frank Coyle

My journey into the mechanics of intelligence and consciousness began during a summer internship at Modesto State Hospital in California. Observing the limitations of traditional psychology in treating severely disturbed patients sparked a four-decade quest to understand how the mind truly functions—and how to better support its complexities.

This pursuit evolved into a rigorous academic path, from psychology at Fordham to the neuroscience labs at Emory. There, I transitioned from dissecting biological brains to discovering the mathematical elegance of McCulloch-Pitts neurons. The arrival of the PDP-8 computer in our lab served as a catalyst; I recognized the same neural patterns I was studying in biological systems mirrored in early computational architecture.

Following a 31-year tenure as a Professor of Computer Science at SMU (where I was known as "Dr. C"), I am now at UC Berkeley, focusing on the frontier of Generative AI and Large Language Models (LLMs). I bring a unique "brain-to-bits" perspective to the classroom, helping students bridge the gap between biological intelligence and modern technology.

My teaching philosophy is guided by the wisdom of avant-garde composer John Cage:

Nothing is a Mistake
There is no win; no lose

Only MAKE!

https://frank-coyle.ai
Previous
Previous

What is agent mean in agentic-AI?