AI Taxonomy
What Is Artificial Intelligence
Artificial Intelligence (AI) studies how to build systems that can perceive environments, represent problems, reason, decide, and act.
From a classical AI perspective, many core problems can be framed as how an agent selects actions in a state space. Search focuses on moving from an initial state to a goal state. Constraint Satisfaction Problems (CSP) focus on finding assignments consistent with variable domains and constraints. Game and multi-agent settings involve strategic decisions in the presence of other agents. Markov Decision Processes and reinforcement learning further unify uncertainty, rewards, and long-horizon decision making.
Machine learning pushes AI from another direction: instead of relying mainly on hand-written rules, models learn patterns from data, feedback, and interaction. Supervised learning learns mappings from labeled data, unsupervised learning discovers structure from unlabeled data, self-supervised learning constructs supervision from data itself, and reinforcement learning learns policies via reward signals.
This page organizes AI knowledge with a modern machine-learning-centric taxonomy. The goal is to provide stable locations for concepts, explain cross-category overlap, and help readers move from isolated terms to connected concept graphs.
Classical AI
This section is reserved for expansion.
Modern Machine Learning and Deep Learning
Detailed taxonomy is maintained in the ML & DL section.
Natural Language Processing
NLP is an independent top-level section: