Watch and track your favorite playlist.
Curated by: Ontology Explained: Philosophy and AI (109 videos)
“How complete is your ontology?” “What does your ontology know about oil and gas?” “Can you quantify it?” These are some of the first questions clients, managers, and stakeholders ask me — and they usually expect a nice clean number like “83%.” But ontology completeness is way more complicated than that. In today’s video, I break down how to properly approach this question, using a metaphor everyone understands: How do you judge what a fifth grader knows? In the video I walk through: 🔹 Why “counting facts” (nodes, edges, depth, connectedness, instance data) only gets you so far 🔹 Why time spent (dev time + application run time) matters — but is also misleading 🔹 Why the real answer is in your tests, applications, and use-case work products 🔹 How competency questions and regression tests map directly to ontology evaluation 🔹 How to pick a strong, narrow, connected use case and build your story around it 🔹 Why you should be suspicious of anyone who claims ontology completeness is simple By the end, you’ll see why evaluating an ontology is a lot like evaluating whether a kid is ready for the next grade: metrics help, but they can be gamed. What really matters is whether the ontology performs across a representative set of tasks. 👇 Let me know in the comments: How do you answer completeness questions? What questions from clients or managers do you struggle with? And hey — the algorithm judges this channel by subscriber count. So help me game that system by liking and subscribing. It helps other people find this content… and it makes me feel good.