Skip to content

jDense — Gnostic Quantifying Weight Dense Layer

Dense layer that applies gnostic quantifying weights (gnostic_weights_j) before the affine transformation.


Overview

jDense mirrors iDense but uses the quantifying weight family (gnostic_weights_j) — computed as the inverse of the estimating weights — before the dense affine transform. It is the quantification-side counterpart to iDense, emphasizing the fj-derived weighting path.


Key Features

  • Applies gnostic_weights_j() (inverse of estimating weights) before dense transform
  • Configurable scale S and lel
  • Full inheritance from Dense

Parameters

Parameter Type Default Description
in_features int Required Input dimension
out_features int Required Output dimension
S float | str 2.0 Scale parameter
lel float None Local Estimate of Location

Methods

forward(x, training=True)

Compute quantifying weights from x, multiply into input, then apply dense affine transform.


Example Usage

from machinegnostics.magnet import jDense
layer = jDense(in_features=2, out_features=1)
print(layer(np.array([[1.0, 2.0]])).shape)

Author: Nirmal Parmar