Neural Loss Landscape Explorer

Landscape & Parameters

Loss Surface Topology

Optimizers Active

SGD Momentum RMSprop Adam

Learning Rate (α) 0.050

Momentum (β1) 0.90

Adam β2 / RMSprop Decay 0.999

Stochastic Noise Variance 0.00

Max Optimization Steps 100

Current Surface Gradient:

∇f = [2x, -2y]

Adam Update Formula:

θₜ₊₁ = θₜ - α / (√v̂ₜ + ε) · m̂ₜ

Click/Drag canvas to set start position: (-1.50, 1.20)

Convergence Telemetry Step: 0

Opt	Loss f(x,y)	Path Len	\|\|∇f\|\|

Loss Decay Curves

Proof Surface Status: READY
Adam Final Loss: -
SGD Final Loss: -