Skip to content

Repository files navigation

Adaptive Intelligence Framework (AIF)

Current research notice: 19 September 2026

AIF studies intelligence as an adaptive boundary-regulation problem. This repository preserves historical AIF manuscripts while the current program is being rebuilt under the SymC General Operations Manual v0.8.0.

Historical AIF v4 remains a research artifact. It should not be read as establishing a universal chi=1 intelligence optimum, a universal 0.8-0.95 adaptive band, or a single scalar law of intelligence.

Current governance

Current migration record: GOM_V0.8.3_MIGRATION_20260921.md.

GOM v0.8.3 requires any scientifically licensed local chi and the broader adaptive-system Chi architecture to be interpreted jointly, while perturbation/recovery remains a separately defined evidence layer. Stability, convergence, consensus, and recovery are not treated as interchangeable concepts.

Current scientific posture

The current AIF direction is architecture-first rather than scalar-first.

Potentially relevant objects include:

  • local and embedded stability;
  • recovery after perturbation;
  • constraint negotiation;
  • modal and multiregime structure;
  • boundary regulation;
  • information/evidence flow;
  • disagreement and diversity preservation;
  • failure/refusal states.

A second-order damping ratio is admissible only when a specific dynamical model earns it. A system is not assigned chi merely because its behavior can be described metaphorically as oscillatory, adaptive, rigid, or chaotic.

Hunting Friction v2

The active Hunting Friction v2 rebuild branch is:

hunting-friction-v2-rebuild

Active unpublished mathematical development, manuscript reconstruction, and protocol work have been moved to the private working workspace.

Public Hunting Friction v2 files on that branch are now status/provenance markers only. Historical public content remains recoverable from Git history.

The retained scientific ceiling is conservative:

  • disagreement reduction is not automatically truth improvement;
  • noninteractive aggregation is a required baseline;
  • first-order consensus is a required null before second-order dynamics are licensed;
  • no universal friction optimum or universal chi optimum is presumed;
  • failure, refusal, and loss of useful diversity are valid outcomes.

Consciousness / adaptive-state work

No new consciousness manuscript or validated consciousness model is currently released from this repository.

Any future consciousness investigation must begin from native neurophysiological/behavioral observables and explicit competing models. It may test whether a stability-oriented architecture adds information beyond established state-space, spectral, arousal/sedation, complexity, and consciousness measures, but it may not define success by recovering a preferred SymC coordinate.

A future public release should follow an explicit prospective freeze and appropriate independent/adversarial validation.

Historical documents

Public historical documents include:

  • SymC_AIFv4.pdf
  • SymC_Methodology.pdf
  • archived AIF bundles

These are preserved for provenance. Later explicit corrections, project-local freezes, and GOM-v0.8.3 controls govern present interpretation where they conflict with older broad language.

Repository status

main is an archival/public landing branch. Current unpublished research lives outside the public manuscript surface until it reaches an evidence-complete release stage.

The program is allowed to discover that a proposed AIF-specific construct adds no value beyond established native methods. That outcome narrows the framework rather than being treated as a failure of process.

About

Adaptive Intelligence Framework (AIF) formalizes intelligence as a boundary-regulated adaptive process. Grounded in SymC principles, it studies how systems maintain coherence under uncertainty through constraint negotiation, regulation, and phase transitions - focusing on adaptive stability rather than benchmark optimization or ML scaling.

Topics

Resources

Stars

3 stars

Watchers

0 watching

Forks

Releases

Packages

Used by

Contributors