AI curriculum

Course outline for Knowledge Engineering Foundations for AI Systems

A model may extract a skill from a profile without reliable evidence that the person has it. Model people, skills, and role requirements as source-qualified claims, concepts, and rules. Compare document retrieval with structured knowledge and keep accepted, inferred, disputed, and unknown information distinct when preparing evidence for AI systems.

Technology: Knowledge Engineering
Duration: 4 hours
Modules: 2 · 2 hours each
Level: Level 1

About Knowledge Engineering Foundations for AI Systems

Knowledge Engineering Foundations for AI Systems Course Objectives

Represent entities and relationships with stable identifiers and defined meanings.
Distinguish taxonomies, ontologies, instance claims, and inference rules.
Compare document retrieval with relationship queries for the same question.
Review supplied model-extracted claims and mark unsupported or conflicting evidence.
Use explicit rules to constrain a mock recommendation or action.
Assign review owners and change triggers while preserving source and conflict history.

Pre-requisites

Familiarity with structured information.
Ability to describe a subject domain and compare conflicting statements; coding is optional.

Before this course

Complete the required courses or bring equivalent experience. Recommended courses add useful depth; they are not enrolment gates.

No prerequisite course is required. The entry skills above still apply.

Can describe a domain and structured information; coding is optional.

Find your place in the learning map

Lab Setup

Diagram editor or spreadsheet and text editor.
Synthetic profiles, source passages, model-extraction fixtures, conflicting claims, and a small rule set.
Retrieval and mock recommendations use supplied artifacts; no coding, live model, vector database, or formal ontology runtime is required.

Detailed Course Outline

Module 1: Make AI evidence and domain meaning explicit

Knowledge representation: Represent people, skills, and requirements as explicit entities and relations; compare document passages with structured evidence for an AI recommendation.
Facts and relationships: Review model-extracted claims against source passages; retain evidence, uncertainty, and conflicting assertions instead of treating fluent output or confidence as verified fact.
Taxonomies: Classify skills with defined labels to improve consistent retrieval; distinguish category membership from instance evidence and inferred relationships.
Ontologies: Define classes and relationship meanings that constrain mock recommendations; separate domain commitments from generated assertions and accepted instance facts.

Practical outcome: Model people, skills, and role requirements; review supplied model-extracted claims against source passages and separate schema, assertions, and uncertainty.

Module 2: Use and steward knowledge for AI systems

Semantic models: Connect shared terms and identifiers; compare a retrieved passage with a structured relationship answer without assuming either automatically resolves missing evidence.
Reasoning: Apply explicit rules to accepted facts, label unresolved answers, and block a mock recommendation when its required supporting evidence is missing.
Knowledge lifecycle: Track source corrections, review owners, and affected downstream AI evidence; retain assertion and conflict history rather than silently replacing uncertain claims.

Practical outcome: Compare document retrieval with structured relationship queries, apply rules to a mock recommendation, and assign source precedence, review owners, and correction triggers.

Practical exercise

Produce a small knowledge model, glossary, source-qualified fact table, and conflict ledger from synthetic profiles and model-extraction fixtures.
Accept a supported skill claim, hold back an unsupported relationship, and record contradictory claims without treating model confidence as evidence.
Answer four questions with known, inferred, disputed, or unknown status; compare a source passage with a structured answer and reject a mock recommendation missing required evidence.
Apply one corrected source record and identify affected facts, rules, and downstream AI evidence while preserving review ownership and history.

What you will produce

Build source-qualified knowledge for AI retrieval and recommendations while keeping claims, rules, uncertainty, and review ownership explicit.

How the work is checked

Schema, instance facts, and rules are clearly separated.
Unsupported and conflicting claims do not become unqualified facts.
Four questions have traceable answers or unresolved status; a mock recommendation with missing required evidence is rejected.
A correction follows the documented review and history rules.

Four-hour noncoding modeling lab with supplied extraction and retrieval fixtures. Provenance records origin, not truth; model confidence does not verify a claim. Teach representation and rules before formal ontology languages, graph implementation, RAG pipelines, or neuro-symbolic integration in later courses.

When another approach is better

A maintained document may suffice when relationships are rarely queried or reused.
Do not freeze an ontology before domain terms and source ownership are understood.

How we train

Ask about exercises, instructor feedback and the prior knowledge you need. Tool-specific courses marked provisional may become modules in a broader course.

Discuss your learning goals