Concept Author: Codrut-Marius Gherasim
R-DKE proposes a machine intelligence architecture capable of instant deep research,
achieved by continuously compressing global knowledge into meaning representations,
verifying truth states, and recursively deepening understanding through autonomous curiosity loops.
This framework outlines:
✅ Semantic knowledge atoms
✅ Truth-graph epistemic memory
✅ Autonomous self-questioning + exploration cycle
✅ Pre-computed reasoning for millisecond answers
✅ Transparent evidence + counter-view synthesis
Try the interactive simulations:
-
Truth Veins vs Noise
Reinforcement of true vs contradictory pathways. -
Living Question Explorer
Autonomous curiosity loop driven by uncertainty. -
Physarum-Inspired Reasoning Loop
Visual ASCII simulation of Physarum-like network growth.
The original design introducing recursive knowledge compression and truth-graph formation.
📄 Full Whitepaper (PDF)
View v1.0 Whitepaper
🧠 Markdown Version
Read v1.0 Markdown
An evolution of R-DKE inspired by Physarum polycephalum,
introducing adaptive reasoning loops where uncertainty acts as nutrient —
allowing knowledge structures to grow, reinforce, or decay organically.
📄 Full Whitepaper (PDF)
View v2.0 Whitepaper
🧠 Markdown Version
Read v2.0 Markdown
A shift from reactive AI → proactive intelligence
“Not a system that waits to be asked.
A system that asks questions first.”
Future work will explore prototypes, evaluation metrics,
and self-regulating safety systems for autonomous reasoning.
| Version | Description | Main Focus | Links |
|---|---|---|---|
| v1.0 | Base Recursive Deep Knowledge Engine | Recursive truth-graph and epistemic compression | Markdown · PDF |
| v2.0 | Physarum-Inspired Adaptive Reasoning | Biological self-organization + uncertainty-driven learning | Markdown · PDF |
Gherasim, M. (2025). Recursive Deep Knowledge Engine (R-DKE).
Physarum-Inspired Recursive Deep Knowledge Engine, v2.0.
License: CC BY 4.0