AI Systems Thinking – Algorithms, Data, Compute, and Knowledge
Lesson 4: Knowledge, Uncertainty, and Feedback
Lesson Objectives
By the end of this lesson, students should be able to:
- Explain knowledge representation in plain language.
- Name uncertainty signals.
- Design a feedback loop.
Lesson Content
AI systems need some way to represent what matters. A rule-based workflow may store conditions. A retrieval workflow may use documents. A recommendation workflow may use course metadata. A planning workflow may use goals, constraints, and steps.
Uncertainty appears when information is missing, instructions conflict, evidence is weak, or the request is outside the workflow. A useful AI system should have a way to slow down: ask for clarification, mark uncertainty, route to a person, or refuse the task.
Feedback helps improve workflows when it is specific. "Bad answer" is less useful than "missed the prerequisite field" or "invented a policy."
Enroll to continue this lesson.
The preview above shows the lesson objectives and opening lesson content. Enroll to view the full lesson, complete the practice work, and take the lesson quiz.