Ontology

The graph is built on two node types: Topics and Skills. Together they form a rich, hierarchical taxonomy of knowledge that the rest of the platform connects to.


Topics and Skills

Topics
Hierarchical knowledge domains

Topics form a taxonomy of human knowledge — from broad fields like “Computer Science” down to specific concepts like “Transformer architecture”. They nest via SUBTOPIC_OF to arbitrary depth.

Topic Types

Field
Subfield
Concept
Method
Technique
Instrument
Genre
Era
Theory
Practice
Tool
Framework
Person
Other
Skills
Abilities and competencies

Skills represent concrete abilities — things a learner can actually do after engaging with content. Every skill relates to at least one Topic viaSKILL_IN_TOPIC.

Technical
Hard skills tied to specific tools, languages, or processes
Soft
Interpersonal and cognitive abilities that transfer across contexts
Psychomotor
Physical skills requiring coordinated body movements
Node Sources
Where Topics and Skills come from

ESCO

European Skills, Competences, Qualifications and Occupations taxonomy — a standardised EU-level skill ontology covering thousands of occupations and skills.

AI-Generated

Topics and skills surfaced by the Librarian agent during cataloging. Reviewed and disambiguated against existing nodes before being committed.

Manual

Nodes created or linked directly by educators, administrators, or learners. The highest-trust source, always honoured for recommendations.

Topics form a taxonomy of human knowledge — from broad fields like Computer Science down to specific concepts like Transformer architecture. Skills represent concrete abilities — things a learner can actually do after engaging with content. Every skill is anchored to at least one Topic via SKILL_IN_TOPIC, so the graph can reason across both domains.


Relationships Within the Ontology

Topics and Skills are not flat lists — they form a rich internal structure:

RelationshipDirectionMeaning
SUBTOPIC_OFTopic → TopicBuilds a hierarchy from broad fields down to specific concepts
SUBSKILL_OFSkill → SkillCreates skill hierarchies, grouping related competencies
SKILL_IN_TOPICSkill → TopicAnchors skills to their domain, enabling cross-domain discovery
RELATED_ESSENTIAL_TOPIC / RELATED_ESSENTIAL_SKILLTopic/Skill → Topic/SkillStrong lateral connections between related areas
RELATED_OPTIONAL_TOPIC / RELATED_OPTIONAL_SKILLTopic/Skill → Topic/SkillWeaker lateral links for broader discovery

Recommendation and discovery logic use these relationships to traverse the graph — for example, finding resources in child topics via SUBTOPIC_OF or in related domains via RELATED_ESSENTIAL_TOPIC.


Where Nodes Come From

The ontology is seeded from ESCO (European Skills, Competences, Qualifications and Occupations), a validated EU taxonomy. New nodes are added when the Librarian agent proposes them during cataloging (after deduplication) or when educators create them manually. Manual and AI-generated nodes are disambiguated against existing ones to avoid fragmentation.


Next Steps