S·ISYNOPTIC.SI
THE ARCHITECTURE OF INTELLIGENCE
← Return to the architecture03 / 09 — FORM

FEATURED / FORM

eigenvalue.si

A measure of transformation.

Some directions change only in scale.

THIS NAMEeigenvalue.si
CONNECTED / 01eigenspace.si↗CONNECTED / 02saddlepoint.si↗CONNECTED / 03statespace.si↗03 / EIGENVALUESELECT A CONNECTION TO MOVE THROUGH THE ATLAS
THE IDEA / 01

An eigenvalue is the factor associated with an eigenvector under a linear transformation.

Where the idea appears

Eigenvalues help describe how a linear transformation scales particular directions, and appear in stability and spectral analyses.

Why it connects

The term pairs with eigenspace to show how a complicated operation can become easier to characterize.

THE QUESTION IT OPENSHow strongly does a transformation act along a direction?
PLACE IN THE ATLASForm

The spaces, transformations, and objectives that shape models.

This is a conceptual collection. The page does not describe a finished product or verified technical capability.
FOLLOW THE THREAD / 02

Connected
by meaning.

Each connection opens a different way to continue through the architecture.

ALSO IN FORM
eigenspace.si ↗losslandscape.si ↗saddlepoint.si ↗symplectic.si ↗cotangent.si ↗
CONTINUE THE ATLAS / ∞

losslandscape.si

Training is a search across possible parameters.

↗