expand_less
Responsible: Ben Goertzel
Papers: Hyperon for AGI⇒ASI Whitepaper (2025), §6.5
Status: Proposed. Concept blending is described in the 2025 whitepaper as a creativity heuristic within PRIMUS. The TransWeave-based formalization and MORK-native implementation are under development. Earlier OpenCog work explored blending in Atomese.
Concept blending is a primary heuristic for creativity — creating novel ideas by merging properties from two or more source concepts so that the result shows useful emergent structure. In Hyperon, blends are not bag-of-words tricks but metagraph constructs that can be tested by the same evaluators that rate patterns, proofs, or programs, making blending a first-class, auditable operation.
How It Works
Blending takes two or more source concepts (represented as subgraphs in AtomSpace), identifies structural alignments between them, and constructs a merged graph that preserves key constraints from each source while introducing novel combinations. The whitepaper gives the intuitive example of "sky-garden" = aerial platform + botanical garden — a blend that inherits structural properties from both sources.
TransWeave Formalization
Under the TransWeave framework, blending becomes a transport path in concept space: choose maps that minimize effort while preserving the constraints each source contributes. Because TransWeave supplies braid/near-commutation rules, blending can be interleaved with pattern mining, inference, or program search and still remain consistent with the system's value geometry.
MORK Implementation
The whitepaper describes an efficient MORK implementation where sources, alignments, and resulting blends are content-addressed — shared sub-parts deduplicate automatically. Transport metadata (who contributed what, with what cost) is stored alongside the blend so future transfers can reuse the same alignments quickly.
Role in PRIMUS
Concept blending operates primarily in PRIMUS's ambient cognitive loop — the background process that maintains and enriches the system's knowledge between goal-directed tasks. Blended concepts become available to:
PLN: new composite hypotheses for inference
Pattern mining: blends may reveal patterns not visible in either source alone
MOSES: blended program fragments as candidate building blocks
Goal management: novel subgoals arising from creative combination
Key References
Goertzel, B. (2025). Hyperon for AGI⇒ASI Whitepaper, §6.5: Concept Blending
Papers: Hyperon for AGI⇒ASI Whitepaper (2025), §6.5
Status: Proposed. Concept blending is described in the 2025 whitepaper as a creativity heuristic within PRIMUS. The TransWeave-based formalization and MORK-native implementation are under development. Earlier OpenCog work explored blending in Atomese.
Concept blending is a primary heuristic for creativity — creating novel ideas by merging properties from two or more source concepts so that the result shows useful emergent structure. In Hyperon, blends are not bag-of-words tricks but metagraph constructs that can be tested by the same evaluators that rate patterns, proofs, or programs, making blending a first-class, auditable operation.
How It Works
Blending takes two or more source concepts (represented as subgraphs in AtomSpace), identifies structural alignments between them, and constructs a merged graph that preserves key constraints from each source while introducing novel combinations. The whitepaper gives the intuitive example of "sky-garden" = aerial platform + botanical garden — a blend that inherits structural properties from both sources.
TransWeave Formalization
Under the TransWeave framework, blending becomes a transport path in concept space: choose maps that minimize effort while preserving the constraints each source contributes. Because TransWeave supplies braid/near-commutation rules, blending can be interleaved with pattern mining, inference, or program search and still remain consistent with the system's value geometry.
MORK Implementation
The whitepaper describes an efficient MORK implementation where sources, alignments, and resulting blends are content-addressed — shared sub-parts deduplicate automatically. Transport metadata (who contributed what, with what cost) is stored alongside the blend so future transfers can reuse the same alignments quickly.
Role in PRIMUS
Concept blending operates primarily in PRIMUS's ambient cognitive loop — the background process that maintains and enriches the system's knowledge between goal-directed tasks. Blended concepts become available to:
PLN: new composite hypotheses for inference
Pattern mining: blends may reveal patterns not visible in either source alone
MOSES: blended program fragments as candidate building blocks
Goal management: novel subgoals arising from creative combination
Key References
Goertzel, B. (2025). Hyperon for AGI⇒ASI Whitepaper, §6.5: Concept Blending