Neural ranking models for document retrieval

Abstract Ranking models are the main components of information retrieval systems. Several approaches to ranking based on traditional machine learning algorithms using a set hand-crafted features. Recently, researchers have leveraged deep in retrieval. These trained end-to-end extract features from raw data for tasks, so that they overcome limitations A variety been proposed, and each model presents neural network used ranking. In this paper, we compare proposed literature along different dimensions order understand major contributions model. our discussion literature, analyze promising components, propose future research directions. We also show analogy between document other tasks where items be ranked structured documents, answers, images videos.