What is symbolic AI?

Represent the problem with human-readable symbols, rules, and states — then reason or search.

Symbolic AI builds an explicit model you can read: facts, Boolean variables, if–then rules, state graphs. Solving means logical deduction or searching for a path to a goal — not only fitting a neural net on pixels.

Symbolic AI

Create a symbolic (human-understandable) representation of a problem and solve it with logic and/or search.

Route through a building

Rooms = states, doors = actions. You do not “learn” the floor plan from photos first — you write the graph and search for a path from entrance to exit. That is symbolic search.

Key takeaways

Symbols + rules + search.
States should be interpretable.
Complements learning-based AI.