# Weighted Global Normalization for Multiple Choice Reading Comprehension over Long Documents

## TL;DR

This paper introduces a weighted global normalization method to enhance answer selection in multiple choice reading comprehension over long documents, significantly improving performance on the NarrativeQA dataset.

## Contribution

The paper proposes a novel weighted global normalization technique for answer selection in long document comprehension tasks, addressing model fragility issues.

## Key findings

- +36.2 Mean Reciprocal Rank improvement
- Effective on NarrativeQA dataset
- Enhances span prediction models for answer selection

## 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.