Interpolated Density Entropy | Neural Layer Demo

Configuration

Optimization Target

Target Entropy 2.0

Point Count (N) 20

Custom Potential

Add to Loss (JS/TFJS)

Evaluated & summed. Vars: x, p (0-1 pos), tf.

Hyperparameters

Temperature (τ) 1.0

Learning Rate 0.01

H = -∫ p(x) log p(x) dx p(x) ∝ 1 / (s_{i+1} - s_i)

Current Entropy --

Loss --

Step 0