An Alternative Cross Entropy Loss for Learning-to-Rank

Sebastian Bruch

Listwise learning-to-rank methods form a powerful class of ranking algorithms that are widely adopted in applications such as information retrieval. These algorithms learn to rank…