Symbolic AI

Symbolic AI

An approach to AI, dominant from the 1950s to the 1980s and sometimes called "Good Old-Fashioned AI" (GOFAI), that represents knowledge explicitly as symbols and rules, then manipulates them through logical inference, rather than learning patterns from data. Expert systems are the best-known application of symbolic AI. It lost ground to connectionism because hand-coded rules struggled to scale to the messiness of real-world data, though hybrid "neurosymbolic" approaches remain an active research area. (See also: Connectionism, Expert system)