English

Weighted Global Normalization for Multiple Choice Reading Comprehension over Long Documents

Computation and Language 2021-11-29 v2

Abstract

Motivated by recent evidence pointing out the fragility of high-performing span prediction models, we direct our attention to multiple choice reading comprehension. In particular, this work introduces a novel method for improving answer selection on long documents through weighted global normalization of predictions over portions of the documents. We show that applying our method to a span prediction model adapted for answer selection helps model performance on long summaries from NarrativeQA, a challenging reading comprehension dataset with an answer selection task, and we strongly improve on the task baseline performance by +36.2 Mean Reciprocal Rank.

Keywords

Cite

@article{arxiv.1812.02253,
  title  = {Weighted Global Normalization for Multiple Choice Reading Comprehension over Long Documents},
  author = {Aditi Chaudhary and Bhargavi Paranjape and Michiel de Jong},
  journal= {arXiv preprint arXiv:1812.02253},
  year   = {2021}
}
R2 v1 2026-06-23T06:33:21.375Z