Ei — Gnostic Entropy Activation¶
Entropy-like activation linking fidelity and residual uncertainty. Case i gives 1 - fi; case j gives fj - 1. Value range: case 'i' → [0, 1); case 'j' → [0, ∞).
Overview¶
Exposes MAGNET's entropy-style family with learnable z0 and S. Case 'i' estimates; 'j' quantifies.
The centered coordinate is θ = (x - z0) / S.
Key Features¶
- Two modes (
i/j) - Learnable
z0,S - Links fidelity and entropy
Parameters¶
| Parameter | Type | Default | Description |
|---|---|---|---|
learnable_S |
bool |
True |
Optimize scale |
learnable_z0 |
bool |
True |
Optimize center |
initial_S |
float |
1.0 |
Initial scale |
initial_z0 |
float \| None |
None |
Initial center |
case |
str |
'i' |
Mode (i / j) |
name |
str \| None |
None |
Name |
verbose |
bool |
False |
Logging |
Attributes¶
S: Stored raw scale parameter tensorz0: Learnable center tensortheta: Most recent centered deviation (θ)case: Active entropy formulation (i/j)last_output: Most recent entropy response
Methods¶
forward(x, training=True)¶
Return the activation applied to input.
- x:
Tensoror array-like — Input values of any shape - Returns:
Tensorwith same shape asx
Example Usage¶
import numpy as np
from machinegnostics.magnet import Ei
activation = Ei()
result = activation(np.array([[0.0, 1.0]]))
print(result.data.tolist()) # [[1.0, 0.0]]
Author: Nirmal Parmar