expand_less
Responsible: Ben Goertzel
Papers: Hyperon for AGI⇒ASI Whitepaper (2025), §5.9
Status: Proposed. Framework for certified subgoal management described in the 2025 whitepaper. Theoretical design; not yet implemented.
SubRep (Subgoal Representation) proposes transforming subgoal and option learning into a disciplined, certifiable practice. Its core mechanisms are formal admissionrulesrules, and a co-learned decomposition networknetwork, thatand algebraic residuation — together determinedetermining when a subgoal genuinely serves a larger purpose and when component solutions can be safely composed.
The Problem
Standard reinforcement learning discovers subgoals by shaping a single scalar reward. This approach is brittle: skills overfit to specific reward signals, transfer negatively to related tasks, and cannot be reliably composed. PRIMUS needs subgoals that work across multiple motives, compose safely, and serve both neural controllers and symbolic reasoners.
Core Mechanisms
SubRep's design centers on four elements described in the whitepaper and supporting glossary:
CDS (Cone-Dominant Subgoals): An option is admitted if robust across a family of motives — not just optimal for one reward, but beneficial across a "cone" of related objectives in motive space. This prevents brittle specialization.
PDS (Pareto-Dominant Subgoals): When genuine trade-offs exist between motives, an option is admitted if Pareto-good on a small covering set — meaning no other available option dominates it on all objectives simultaneously.
MDN (Motive Decomposition Network): A co-learned network that decomposes high-level motives into achievable subgoals, mapping motive space into option space. The MDN learns which skills serve which purposes and identifies gaps.
Weakest-sufficientResiduation decomposition:(A* = S/B): SubRep uses algebraic residuation from the weakness framework to findcompute decompositionsthe that"weakest-sufficient" aremissing sufficientpiece forof a plan — the taskminimal butadditional nocapability strongerneeded thanto necessarybridge —the preventinggap between current skills and a target goal. This prevents over-specialized subgoals that fail to transfer.
Each admitted option is associated with a Decision Transformer (To) — a learned policy conditioned on the option's initiation and termination conditions. Certificates are expressed as backed-up values over AtomSpace features, making them native to the broader neurosymbolic loop.
Neurosymbolic Design
Unlike standard RL option frameworks, SubRep is designed to work equally for neural controllers, PLN-derived logic macros, and MOSES/GEO-EVO evolved programs — all screened by the same CDS/PDS admission rules and composed by the same planner.
Relationship to MetaMo and TransWeave
SubRep is designed as a complement to MetaMo: MetaMo defines what the system cares about as motive geometries; SubRep validates which skills serve those motives with formal certificates. When TransWeave transfers skills, SubRep certificates would travel with them.
Key References
Goertzel, B. (2025). Hyperon for AGI⇒ASI Whitepaper, §5.9: SubRep
Papers: Hyperon for AGI⇒ASI Whitepaper (2025), §5.9
Status: Proposed. Framework for certified subgoal management described in the 2025 whitepaper. Theoretical design; not yet implemented.
SubRep (Subgoal Representation) proposes transforming subgoal and option learning into a disciplined, certifiable practice. Its core mechanisms are formal admission
The Problem
Standard reinforcement learning discovers subgoals by shaping a single scalar reward. This approach is brittle: skills overfit to specific reward signals, transfer negatively to related tasks, and cannot be reliably composed. PRIMUS needs subgoals that work across multiple motives, compose safely, and serve both neural controllers and symbolic reasoners.
Core Mechanisms
SubRep's design centers on four elements described in the whitepaper and supporting glossary:
CDS (Cone-Dominant Subgoals): An option is admitted if robust across a family of motives — not just optimal for one reward, but beneficial across a "cone" of related objectives in motive space. This prevents brittle specialization.
PDS (Pareto-Dominant Subgoals): When genuine trade-offs exist between motives, an option is admitted if Pareto-good on a small covering set — meaning no other available option dominates it on all objectives simultaneously.
MDN (Motive Decomposition Network): A co-learned network that decomposes high-level motives into achievable subgoals, mapping motive space into option space. The MDN learns which skills serve which purposes and identifies gaps.
Each admitted option is associated with a Decision Transformer (To) — a learned policy conditioned on the option's initiation and termination conditions. Certificates are expressed as backed-up values over AtomSpace features, making them native to the broader neurosymbolic loop.
Neurosymbolic Design
Unlike standard RL option frameworks, SubRep is designed to work equally for neural controllers, PLN-derived logic macros, and MOSES/GEO-EVO evolved programs — all screened by the same CDS/PDS admission rules and composed by the same planner.
Relationship to MetaMo and TransWeave
SubRep is designed as a complement to MetaMo: MetaMo defines what the system cares about as motive geometries; SubRep validates which skills serve those motives with formal certificates. When TransWeave transfers skills, SubRep certificates would travel with them.
Key References
Goertzel, B. (2025). Hyperon for AGI⇒ASI Whitepaper, §5.9: SubRep