Principle:Mbzuai oryx Awesome LLM Post training Category Taxonomy Definition
| Knowledge Sources | |
|---|---|
| Domains | Curation, Information_Architecture |
| Last Updated | 2026-02-08 07:30 GMT |
Overview
An editorial process that defines a hierarchical category structure for organizing curated resources based on a survey paper's taxonomy.
Description
Category Taxonomy Definition establishes the organizational skeleton of a curated awesome list. The taxonomy is derived from a companion survey paper's topic hierarchy and translated into markdown section headers. This creates a navigable structure where each top-level section represents a major research area and subsections capture specific topics within it.
The quality of the taxonomy directly determines the usability of the curated list: too broad and papers are hard to find, too granular and the structure becomes fragmented. The taxonomy should align with how the research community organizes the field.
Usage
Use this principle when:
- Creating a curated resource list (awesome list) for a research domain
- A survey paper or authoritative source provides a field taxonomy
- The resource collection is large enough to warrant categorical organization
- The taxonomy needs a table of contents for navigation
Theoretical Basis
The taxonomy definition process follows information architecture principles:
Pseudo-code Logic:
# Abstract taxonomy definition pattern (NOT real implementation)
taxonomy = derive_from_survey_paper(
source="companion_survey.pdf",
figure="taxonomy_figure"
)
sections = []
for category in taxonomy.top_level:
section = create_section(level=2, name=category)
for subcategory in category.children:
subsection = create_section(level=3, name=subcategory)
section.add(subsection)
sections.append(section)
table_of_contents = generate_toc(sections)
Design criteria:
- Mutual exclusivity: Categories should minimize overlap
- Collective exhaustiveness: Every relevant paper should fit in at least one category
- Navigability: Table of contents with anchor links