# CHARM Quark

**One-liner**  
A noise-resilient RL model that preserves coherence when signal-to-noise collapses.

## Why it matters

In high-noise environments, most agents either freeze or overfit noise. CHARM maintains usable structure.

* Environment: High stochasticity
* Signal: Coherence decay
* Mechanism: Noise-adaptive filtering
* Outcome: Stable learning under chaos

## What makes it different

* Noise as first-class input
* Coherence preservation
* Anti-overfitting bias

## Status

Experimental research prototype.