Calvin Lab Auditorium
We discuss a recent line of work on classical linear algebra algorithms inspired by quantum machine learning. Such algorithms disprove that their corresponding quantum algorithms admit exponential speedups. Further, they show an interesting connection between QML and classical sketching and sampling literature, which helps clarify where exactly quantum algorithms gain advantage over classical. We discuss some lemmas central to this line of work, algorithms based on these lemmas, and some open questions in this space.