> but practical algorithms with better exponents exist for all of them. I'm aware of randomized algorithms with better complexity, which come at the cost of on

> but practical algorithms with better exponents exist for all of them. I'm aware of randomized algorithms with better complexity, which come at the cost of only giving approximate results (though the approximation may be perfectly good for practical purposes). See e.g. [1]. Are there other approaches? [1] https://doi.org/10.1137/090771806 https://doi.org/10.1137/090771806