Evolution meets learning.

Same seeded course. Evolution versus learning with replay.

Sim Speed (1x)
🧬 GENETIC ALGORITHM (POP: 20)
📉 GRADIENT DESCENT (ONLINE BACKPROP)
REPLAY BUFFER: 0

Live metrics Gen: 1

Genetic Algorithm

Alive: 20
Best Score: 0
Splits Left: 4

Gradient Descent

Loss (MSE): 0.0000
Best Score: 0
Updates: 0

GA: Explores architectures via population splits & crossover.

GD: Adjusts weights via Backprop & Replay. Accepts weights on success, reverts on hit (Epigenetic Gating).

Simulation Event Log Tracking Heredity & Replay Updates