Singular Value Decomposition (SVD) - A matrix factorization technique that decomposes a matrix into three simple matrices, revealing geometric structure and enabling tasks like compression, noise reduction, and solving linear systems.