1: \begin{abstract}

2: We present a new perspective on the popular Sinkhorn algorithm, showing that it can be seen as a Bregman gradient descent (mirror descent) of a relative entropy (Kullback--Leibler divergence). This viewpoint implies a new sublinear convergence rate with a robust constant.

3: \end{abstract}

4: