Thanks to visit codestin.com
Credit goes to Github.com

Skip to content

MercAIA/C7

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 

Repository files navigation

C7 Architecture

C7 is a dual-hemisphere cognitive architecture designed to move beyond classic pattern-matching AI systems. It introduces a self-regulating, multi-state reasoning core that dynamically switches between shallow and deep computation based on internal signals such as:

  • predicted error
  • coherence signals
  • input intensity
  • surprise response
  • feedback sensitivity
  • grounding stability

It was developed through iterative experiments from multimodal front-ends (audio, text, image) to a unified A7 integrator with a gated shallow/deep processor.


Features

  • Emb-C (collapsed multimodal embedding)
  • Intensity estimation (audio/text/image energy)
  • Associative arrays A1, A3, A5
  • A7 integrator with stability regulation
  • Shallow/Deep dual-mode processing
  • Surprise-triggered gating
  • Grounding layer for architectural stability
  • Adaptive reasoning effort
  • Experimental training loops

Repository Structure

/core
c7_core_v1.py
c7_core_training.py

/modules
audio_frontend.py
text_frontend.py
image_frontend.py
emb_c.py
integration_a7.py
gating.py

/experiments
phase1_to_phase9/
surprise_gate/
intensity_tests/
coherence_tests/

/docs
whitepaper/
diagrams/


Whitepaper

C7: Two-Hemisphere Grounded Intelligence
DOI: 10.5281/zenodo.17640165


Roadmap

  • add datasets for real multimodal training
  • build AudioBrain (C7-A)
  • extend grounding layer
  • create neuroscience-aligned version (C7-Neuro)
  • open-source demo checkpoint
  • release trained model v1.1

License

Open for academic research.
Commercial licensing via MercAIA.

Contact

MercAIA Project
Mostafa Bahram