Making sense of AI taxonomy

6 min read · Jun 7, 2026

I have been looking for diagrams that classify AI types, methods, and definitions. It feels like no one agrees on the definitions. Every diagram differs from one source to another. Instead of helping me understand the field as a whole, it confused me even more.

I decided to make my own, with Wikipedia, React Flow, and time.

I am certain there are many things I have missed, but this was my first step toward understanding the big picture. If you spot any mistakes, please feel free to contact me at me@kaonashi.dev.

Some thoughts after doing this exercise: AI taxonomy should be a huge graph with connections going in all directions, or lots and lots of smaller diagrams. Here I mixed paradigms, architectures, models, learning methods and so on so I could make a mental map of it all.

You may click on boxes to get a definition.

AI Taxonomy

Artificial Intelligence

Symbolic AI

Machine Learning

Hybrid AI

Logic Programming

Knowledge Graph Reasoning

Expert Systems

Supervised Learning

Unsupervised Learning

Reinforcement Learning

Deep Learning

Neural Networks

MLP

FNN

CNN

RNN

LSTM

GRU

Transformer

GNN

Autoencoder

GAN

Diffusion Model

Neuro-symbolic AI

AI Systems

LLM-based Chat Systems

RAG Systems

Agent Systems

Multi-component AI Pipelines

Large Language Models

Encoder-only

BERT

Decoder-only (LLMs)

GPT

LLaMA family

Mistral AI

Gemma

Claude

Encoder-Decoder

BART

React Flow

Node Types

paradigm

learning

architecture

model family

generative model families

system

method

system composition

transformer family

model