Minimax Regret for Online Learning with Feedback Graphs

The dataset used in the paper is a sequence of strongly observable undirected feedback graphs, where each graph has independence number α for some common value α.

BibTex: