Bottom-up taxonomy
Organizing information by observing many small examples and grouping them into categories, rather than starting with predefined ones.
Imagine you have a big box of LEGO bricks. Instead of deciding beforehand you want to build a car, a house, and a spaceship, you start by looking at all the bricks. You might notice some are red, some are blue, some are big, some are small, and some have wheels. You start sorting them into piles based on these observations: all the red bricks together, all the wheels together, all the big bricks together.
This is like a bottom-up taxonomy. You build your categories by looking at the actual pieces (or in AI, the actual uses of AI) and finding patterns. You might end up with categories like 'bricks for building walls' or 'pieces for making things roll,' which you didn't plan at the start. It's about letting the data guide the organization.
A bottom-up taxonomy is a classification system derived from empirical observation of instances, where categories emerge from the data itself rather than being imposed by a predefined structure. This approach contrasts with top-down methods that start with broad, abstract categories and then subdivide them. In the context of AI applications, a bottom-up taxonomy involves analyzing usage logs, user queries, or documented AI outputs to identify recurring patterns and functionalities, which are then grouped into meaningful categories. This method is useful for discovering emergent use cases and understanding the practical landscape of AI deployment, but it can be more labor-intensive and may result in a less immediately intuitive structure compared to a top-down approach.
Anthropic's recent economic index reports highlight a granular taxonomy of AI applications. This bottom-up approach allows for a detailed understanding of how AI is being used across various sectors, moving beyond broad categories to specific instances of application.
- Anthropic Economic Index: Insights from Claude 3.7 SonnetAnthropic · Anthropic Research ·
- Anthropic Economic Index: Insights from Claude 3.7 SonnetAnthropic · Anthropic News ·
- Anthropic Claude
- Anthropic uses a bottom-up taxonomy to categorize and analyze the specific applications and economic impact of its Claude models.
- Top-down taxonomy
- A classification system that starts with broad, predefined ca
- Folksonomy
- A user-generated tagging system where individuals categorize content
- Controlled vocabulary
- A curated list of terms used for indexing and retrieval,
The concept of bottom-up classification originates from fields like biology and library science, where systems were developed by observing and grouping organisms or documents based on shared characteristics. In computer science and AI, it's applied to data analysis and knowledge representation.
Choose a large text document or a collection of short articles. Read through them and start highlighting recurring themes or topics. Group these highlighted sections into categories based on what you observe. For example, if you're reading news articles, you might start seeing patterns around 'politics,' 'technology,' and 'sports' without having decided those categories beforehand.
The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.
- PromptRole-play a student struggling with math for AI tutor simulationSimulate a student's math struggles and uncertainty for AI tutor evaluation. · 5 min
- TemplateReusable workflow templates for healthcare teamsTeams can create and share templates for common workflows like documentation or prior authorizations. · 15 min
- TemplateTemplate for building an AI tool tailored to your workA structured approach to defining and building a custom AI tool for specific business needs. · 15 min
- TemplateEvals CSV prompt list for agent skillsA CSV file template for listing prompts used in evaluating agent skills. · 20 min
- HeuristicDouble coding agent task completion with docs linksProviding coding agents with a documentation link significantly increases their task completion rate. · 2 min
Questions people ask
- What is the main advantage of a bottom-up taxonomy?
- The main advantage is that it reflects the actual data and emergent patterns, potentially uncovering categories or relationships that might be missed by a predefined system.
- When would a top-down taxonomy be better?
- A top-down taxonomy is often better when you have a clear, established framework or when you need a highly organized and predictable structure from the outset.
First seen on Subvolts Aug 25, 2026. Updated Sep 18, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.