We present algorithms with optimal sample complexity and running time for estimating a high dimensional Euclidean mean.

We theoretically characterize MoM's optimality for different classes of distributions under adversarial contamination