Document and Query Likelihood Evaluations for Learning to Rank Systems

Assessing Quality of Neural Re-ranking based on certain training objectives