We study replicable distribution testing.

We demonstrate adversarially replicable online learning algorithms for online linear optimization and the experts problem that achieve sub-linear regret.