Passage Filtering for Open-Domain Question Answering

摘要

This paper studies the problem of filtering data in Passage Retrieval applied to Question Answering. Specifically, in this paper we have proved that the Mean-Value Theorem can play an important role to improve Question Answering. We have studied the way in which this theorem can be applied in order to produce a maximum data reduction without precision loss. In the experiments, we achieve a 90% data reduction without significant data loss.