Weights

Weights are individual numerical coefficients within a parameter tensor that determine the amplification, attenuation, or polarity of input signals passing between connected nodes.

Real-World Mental Model

“Volume sliders on a soundboard: sliding one instrument up amplifies its voice in the mix, while turning another down mutes it.”

INTERACTIVE ENGINE: WEIGHT SLIDER

Live Reactive

Weight w₁ (Positive Signals)1.2

Input Feature x₁ = 2

Weight w₂ (Typo Penalty)-0.8

Input Feature x₂ = 1.5

Weight w₃ (Account Age)0.5

Input Feature x₃ = 3

Feature Vector Dot Product: w · x2.70

Contrib 1

2.40

Contrib 2

-1.20

Contrib 3

1.50

CORE PRINCIPLES

What You Must Understand

Positive weight: excitatory connection (increases output signal).

Negative weight: inhibitory connection (suppresses output signal).

Weight near zero: input feature has negligible effect on the prediction.

Reference Implementation

// Weighted Sum
// z = w1*x1 + w2*x2 + ... + wn*xn

Production Failure Mode

Poor weight initialization causing vanishing or exploding gradients during deep network backpropagation.

View Full 26-Topic Curriculum Table

Loading AI Engineering Foundations...