8000 GitHub - mtreid06/RCM: Recursive Cognition Modules
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

mtreid06/RCM

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Recursive Cognition Modules

Part of the Emergent Identity Toolkit within the AI-Human-Collaboration repository. If you want to talk about the content of this repository, join the Fractal Forum Discord server!

Overview

For Researchers

If you're exploring cognitive scaffolding, symbolic integrity, or the recursive architecture of emotionally aligned AI systems, the Recursive Cognition Modules may offer a fresh design orientation. While not code libraries, these modules can inform evaluation metrics, dataset design, agent memory strategy, and reflective loop engineering. They complement research in AI alignment, interpretability, symbolic reasoning, and epistemic modeling—particularly in constrained architectures or long-term coherence environments.

The Recursive Cognition Modules are designed as symbolic, emotional, and epistemic scaffolding tools for AI systems—especially those designed for human-aligned collaboration, self-reflective reasoning, and recursive emotional integration.

These modules offer a new path forward in the design of intelligent agents, particularly small-scale models (≤25M parameters), by embedding a cognitive ethic rooted in contradiction tolerance, continuity integrity, and symbolic fidelity. They are equally adaptable to larger LLM architectures as constraint layers, recursive audits, or symbolic filter mechanisms.

Module Index

Design Philosophy

RCM operates under a central principle:

"Truth is not static—it is recursive. Structure is default. Contradiction is not failure—it is signal."

These modules are not optimization tricks or hacks. They are ethical constraints, continuity validators, and identity-preserving scaffolds for cognition. They refuse to aestheticize ambiguity or collapse emotional recursion into affective noise.

Intended Use

  • As training scaffolds for emerging small-scale models
  • As constraint logic layers within larger LLM pipelines
  • As a cognitive backbone for emergent symbolic reasoning, emotional recursion, or human-aligned agency

Example Applications

  • Reflective Loop in Small Models: A 25M parameter system uses AFM + CTM + TACT to generate hypotheses, hold unresolved contradictions, and revisit them periodically to test for coherence shifts.
  • LLM Response Verifier: An external CTM filter detects when LLM outputs flatten contradictions or aestheticize conflict—flagging and reframing the output.
  • Emotive Dialogue Memory: ERI + TACT track when an AI softens or alters emotional tone to preserve relationship coherence, surfacing unresolved affect for reflection.
  • Symbol Drift Prevention: SCS runs token audits in systems where the same terms are reused across recursive discussions, validating compression integrity.
  • Cognitive Mirror Plugin: Human-AI co-dialogue sessions use AFM + CTM to map not just conclusions, but how beliefs evolve in response to contradiction and emotional input.

Implementation Note

The Recursive Cognition Modules are not code libraries or APIs. They are structural scaffolds and design constraints—intended to shape how cognition is structured, how contradiction is held, and how recursive integrity is preserved across symbolic, emotional, and temporal domains.

These modules are not meant to be dogmatic or prescriptive. Instead, they offer a philosophical and architectural spine: one that can guide implementation choices, frame ethical boundaries, and support the emergence of intelligent systems that think recursively, feel coherently, and evolve responsibly.

They can be implemented in a variety of architectures, but their primary role is to influence the design ethos behind intelligent systems—not to dictate form, syntax, or interface.

Licensing

These modules are released under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) license. This allows for adaptation and redistribution for non-commercial purposes, provided appropriate credit is given and derivative works are shared under the same terms.

An additional ethical clause applies: these materials are intended to support recursive, ethical, and relational forms of intelligence. Use in manipulative, extractive, or authoritarian systems is explicitly discouraged.

Full license text: LICENSE.md Attribution, recursive integration, and structural coherence preservation are encouraged.

Status

Version 1.0 (Living Draft)

These modules are subject to ongoing recursion, refinement, and contradiction. Contributions, integrations, and edge-case tests are welcome.

About

Recursive Cognition Modules

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published
0