Linear Algebra

This lecture explores the fundamental geometric interpretations of linear equations, delving into row pictures and column pictures to visualize solution sets and vector spaces in two and three dimensions.

Turns Gaussian elimination from tedious bookkeeping into matrix algebra by deriving elimination and row-exchange matrices from first principles.

The rank-1 outer-product view of matrix multiplication taught at 16:00 is essential preparation for SVD, PCA, and low-rank adaptation (LoRA) in ML.

Shows why inverse elementary matrices assemble without cross-terms into lower-triangular L, yielding A = LU and quantifying the exact (1/3)n³ operation count of Gaussian elimination.

Marks the transition from row operations to genuine vector-space geometry: symmetric matrices RᵀR, P⁻¹ = Pᵀ for permutation matrices, subspaces of ℝⁿ, and closure under linear combinations.

Introduces the first two of Strang's Four Fundamental Subspaces—connecting solvability of Ax = b to C(A) and uniqueness of solutions to N(A).

Systematic algorithm for computing a basis of the nullspace N(A) from echelon form, proving why dim N(A) = n - r free variables.

Unifies linear systems completely: x_complete = x_particular + x_nullspace, classifying full column rank, full row rank,full invertible rank, and rank-deficient matrices.

This session rigorously defines linear independence, spanning sets, and basis vectors, culminating in the formal definition of dimension for vector spaces and subspaces.

The lecture introduces the crucial concepts of the column space, null space, row space, and left null space of a matrix, elucidating their interrelationships and orthogonal properties.

This lecture examines different types of matrix spaces, with a focus on rank-1 matrices and their construction. It also introduces the concept of small-world graphs as an application.

This session explores the connection between linear algebra and graph theory, demonstrating how incidence matrices represent graphs and their properties in network analysis.

This lecture introduces the concept of orthogonal vectors and subspaces, establishing the conditions for orthogonality and its significance in vector space decomposition.

The session details the process of projecting a vector onto a subspace, deriving the projection formula and its geometric interpretation in finding the closest point.

This lecture explores the construction and properties of projection matrices, demonstrating their application in solving overdetermined systems via the method of least squares.

This session introduces orthogonal matrices and their properties, followed by a detailed explanation of the Gram-Schmidt orthonormalization process for constructing orthonormal bases.

This lecture systematically covers the axiomatic properties of determinants, including their behavior under row operations, transposition, and scalar multiplication.

This session derives explicit formulas for determinants using permutations and introduces the cofactor expansion method for calculating determinants of arbitrary size matrices.

This lecture presents Cramer's Rule for solving linear systems, connects the adjoint matrix to the inverse, and illustrates the geometric interpretation of determinants as volume.

This foundational lecture introduces the concepts of eigenvalues and eigenvectors, demonstrating their significance in understanding matrix transformations and system dynamics.

Widely cited across Hacker News and math forums as the single clearest opening lecture in undergraduate mathematics: reframes systems of equations as linear combinations of column vectors within 15 minutes.