Frame-Based Knowledge Representation

Origin. Minsky's frames (1974) were extensively developed in Soviet AI at the Institute of Control Sciences and elsewhere; integration with Soviet logic programming and semantic networks.

Mechanism. Knowledge is represented as frames — structured records with slots for attributes, values, and relations. Frames inherit from parent frames, support default values, and trigger procedures when accessed (demons). The representation is closer to human conceptual structure than predicate logic, supporting typical-case reasoning and exception handling. Frames can represent both classes and instances, taxonomic and compositional relations.

Procedure. Model the domain as a hierarchy of frames. Define slots for properties, with default values and constraints. Define is-a relations for inheritance and part-of relations for composition. Attach demons to slots: if-needed demons compute values lazily, if-added demons maintain consistency, if-removed demons handle deletion. Query by pattern matching or slot access. Use inheritance to avoid redundancy; specialize by overriding defaults.

Applies to. Expert systems, knowledge bases, semantic modeling, any domain where structured knowledge with defaults and inheritance matches the domain structure.

Limitations. Multiple inheritance creates ambiguity when ancestors conflict. Default reasoning is non-monotonic: adding information can retract conclusions, which complicates truth maintenance. The frame-logic integration is not fully formalized, so frame systems can behave unexpectedly. Soviet work emphasized the engineering practice over formal semantics.