jgrusewski
|
44b38e8014
|
docs: RL agent gaps design — 7 features for training quality and production readiness
Position-aware state encoding, regime detection (feature-based + HMM),
slippage modeling, curriculum learning, multi-timeframe fusion (default on),
online learning with EWC, and A/B testing with Thompson Sampling.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
|
2026-03-03 02:23:38 +01:00 |
|