Embedded manifolds are smooth subspaces of a higher-dimensional Euclidean space that locally resemble Euclidean space and are defined by smooth embeddings. They enable the study of complex geometries and topologies within the context of more familiar Euclidean spaces, facilitating analysis and computation in fields like differential geometry and machine learning.