Robust Contextual Bandit via the Capped-$\ell_{2}$ norm

Open in new window