For Linear Algebra Gilbert Strang | Lecture Notes

| Section | Content | |---------|---------| | Key insight (1 sentence) | What is the single big idea today? | | Main example | The small matrix or vector space he keeps returning to. | | New definition | In his words, then in your own. | | Connection to the 4 subspaces | Where does today’s topic fit? | | Computation method | Steps for solving/calculating (if any). | | Typical exam question | Predict one. | | Confusion point | Note what you need to rewatch. |

For an (m \times n) matrix (A) (rank (r)), there are four fundamental subspaces:

| Subspace | Notation | Dimension | Contained in | |----------|----------|-----------|---------------| | Column space | (C(A)) | (r) | (\mathbbR^m) | | Nullspace | (N(A)) | (n - r) | (\mathbbR^n) | | Row space | (C(A^T)) | (r) | (\mathbbR^n) | | Left nullspace | (N(A^T)) | (m - r) | (\mathbbR^m) |

Strang’s Fundamental Theorem of Linear Algebra:

The beauty of these lecture notes lies in their universality:

When (Ax = b) has no solution, we solve (A^TA\hatx = A^Tb). This minimizes (|Ax - b|^2). The least squares solution is: [ \hatx = (A^TA)^-1A^T b ] lecture notes for linear algebra gilbert strang

Key geometric insight: The error (e = b - A\hatx) is perpendicular to the column space of (A).

Date: [today]                Topic: Least Squares
Key insight: When Ax=b has no solution, find x̂ that minimizes the error.
Main example:
Points (0,0), (1,1), (2,2) → line b = 0 + 1*t → perfect.
Points (0,0), (1,1), (2,3) → no line through all three.
A = [1 0; 1 1; 1 2], b = [0;1;3]
Normal equations: A^T A x̂ = A^T b
A^T A = [3 3; 3 5], A^T b = [4;7]
Solve: x̂ = [1; 0.5] → line b = 1 + 0.5 t
Connection to 4 subspaces:
Error e = b - A x̂ is perpendicular to C(A)
So e is in N(A^T) | Section | Content | |---------|---------| | Key
Computation method:

Typical exam question:
Find least squares line through (-1,0), (0,1), (1,2).
Confusion point:
Why (A^T A) invertible? → When A has independent columns.

To give you a taste of what high-quality lecture notes for linear algebra Gilbert Strang look like, here is a condensed summary of the most critical lecture: Typical exam question: Find least squares line through

Lecture 10: The Four Fundamental Subspaces

For an ( m \times n ) matrix ( A ) of rank ( r ):

The Big Picture: ( N(A) ) is orthogonal to ( C(A^T) ). ( N(A^T) ) is orthogonal to ( C(A) ).

A full set of notes would then show you why the rank reveals the dimension of each space and how elimination exposes their bases.

Unlike many traditional mathematics courses that prioritize rigorous proof over concept, Gilbert Strang’s notes are built on a philosophy of visual intuition. The notes do not begin with abstract definitions of vector spaces; they begin with the fundamental problem: $Ax = b$.

The notes are famous for de-emphasizing the tedious calculation of determinants (often relegated to the latter half of the course) and prioritizing the Column Space and Eigenvalues. Strang’s central teaching philosophy is that "linear algebra is the study of vectors and matrices." His notes focus on seeing the "big picture"—visualizing vectors moving in space, understanding matrices as operators that transform that space, and grasping the geometry behind the algebra.