Given a matrix $A \in \mathbb{R}^{n\times n}$ and $m\in \mathbb{N}$, I'd like to find a matrix $B \in \mathbb{R}^{n\times m}$, such that
$$B B^\top \approx A$$
where $\approx$ is intentionally vague and can mean anything practical like "in the least squares sense" or something similar.
Does this problem have a name and simple way of solving it?