Reading series · 8 essays

Geometry → Representation → ML

From coordinates and projections to semantic spaces and statistical manifolds.

  1. 1Part 1

    Basis: Choosing Coordinates for a World

    A basis is not the vector space; it is the coordinate language we choose to describe it.

  2. 2Part 2

    Projection: The Shadow of a Vector

    Orthogonal projection turns the intuitive idea of a shadow into one of linear algebra's most useful operations.

  3. 3Part 3

    Cosine Similarity: When Angle Becomes Meaning

    Normalize the dot product and vector angle becomes a practical measure of similarity.

  4. 4Part 4

    Linear Transformations: Functions That Respect Geometry

    See matrices as machines that transform entire grids rather than as rectangular tables of numbers.

  5. 5Part 5

    Eigenvectors: Directions That Refuse to Turn

    Some directions survive a linear transformation without changing direction. Those are eigenvectors.

  6. 6Part 6

    Covariance: How Variables Move Together

    From scatter clouds to covariance ellipses, principal directions and positive semidefinite matrices.

  7. 7Part 7

    Word2Vec: When Words Acquire Geometry

    How prediction turns words into points, neighborhoods and surprisingly useful directions.

  8. 8Part 8

    Information Geometry: When Probability Distributions Become Points

    A first walk through statistical manifolds, Fisher information and the geometry hidden inside probability.