Samuel Tenka

$\newcommand{\NN}{\mathbb{N}}$ $\newcommand{\ZZ}{\mathbb{Z}}$ $\newcommand{\RR}{\mathbb{R}}$ $\newcommand{\CC}{\mathbb{C}}$ $\newcommand{\pr}{^\prime}$ $\newcommand{\prpr}{^{\prime\prime}}$ $\newcommand{\wrap}[1]{\left(#1\right)}$

Phrasebook for Math Modeling

These notes, broken into 24 lessons, are a math curriculum for very motivated folks who wanna get into physics, machine learning, etc but don't have special math background beyond highschool math plus intellectual maturity. By intellectual maturity I mean that invaluable bag of attitudes and understandings that recognizes how names and named are different, that learns general patterns from specific examples, that values "what if"s, etc.

Topic-wise, it's my reworking of the actually practical parts of college math. Whereas college math in my country typically focuses a lot on Calculus, I think that concepts relating to Probability and Linear maps are more important, for a well-educated-hence-happier citizenry, than most of the calculus details in current curricula. So here we have 4 lessons each on: Probability, then Calculus, then Linear Maps; we then mix these ideas when examining Geometry in the Small and Uncertainty about the Infinite; we end with a grab bag of Special topics that illustrate how these math ideas help us analyze interesting problems.

At a finer grain, here's what's inside:

Grades of belief ; Bayes law and independence ; Expectations and abstraction ; Application: minimax game theory ; Limits, Derivatives ; Integration, DiffEQs, Exp ; Texture of the Continuum: Taming Pathology ; Taylor Series and Asymptotics ; Maps as Matrices ; Dimension and Determinant ; Vectors vs Covectors ; Inner Products and Geometry ; Taylor in Multiple Variables ; Densities and Integration ; Vector Fields ; Gaussians and Spheres ; Probability Densities ; Central limit theorem ; Theory of measurement: statistics and error propagation ; Statistical manifolds ; Fourier analysis and oscillations ; Control theory and optimization ; Regression and smooth models ; Energy-based models and statmech

Grades of belief

Credence levels as numbers
Probability theory is a math encoding of some of our

The idea is that we have a bunch of statements --- each apt to be true-or-false (rather than a command, or question, or super-vague) --- and we want to put a number on each statement that says how much we believe it.

Bayes law and independence

Expectations and abstraction

Application: minimax game theory

Limits, Derivatives

Approximation

Integration, DiffEQs, Exp

Texture of the Continuum: Taming Pathology

Taylor Series and Asymptotics

Maps as Matrices

Dimension and Determinant

Vectors vs Covectors

Inner Products and Geometry

Taylor in Multiple Variables

Densities and Integration

Vector Fields

Gaussians and Spheres

Probability Densities

Central limit theorem

Theory of measurement: statistics and error propagation

Statistical manifolds

Fourier analysis and oscillations

Control theory and optimization

Regression and smooth models

Energy-based models and statmech

References

Author --- Title --- Year