# Improving Machine Reading Comprehension with General Reading Strategies

arXiv: https://arxiv.org/abs/1810.13441 · HTML: https://mlsanity.com/paper/1810.13441/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 31 Oct 2018.

## Built on by (most cited first, 20 of 20)

- [Learn to Resolve Conversational Dependency: A Consistency Training Framework for Conversational Question Answering](https://mlsanity.com/paper/2106.11575/) (2021)
- [What Ingredients Make for an Effective Crowdsourcing Protocol for Difficult NLU Data Collection Tasks?](https://mlsanity.com/paper/2106.00794/) (2021)
- [DREAM: Improving Situational QA by First Elaborating the Situation](https://mlsanity.com/paper/2112.08656/) (2021)
- [Narrative Question Answering with Cutting-Edge Open-Domain QA Techniques: A Comprehensive Study](https://mlsanity.com/paper/2106.03826/) (2021)
- [Salience-Aware Event Chain Modeling for Narrative Understanding](https://mlsanity.com/paper/2109.10475/) (2021)
- [Graph-free Multi-hop Reading Comprehension: A Select-to-Guide Strategy](https://mlsanity.com/paper/2107.11823/) (2021)
- [Clues Before Answers: Generation-Enhanced Multiple-Choice QA](https://mlsanity.com/paper/2205.00274/) (2022)
- [Back to the Future: Bidirectional Information Decoupling Network for Multi-turn Dialogue Modeling](https://mlsanity.com/paper/2204.08152/) (2022)
- [Self-Teaching Machines to Read and Comprehend with Large-Scale Multi-Subject Question-Answering Data](https://mlsanity.com/paper/2102.01226/) (2021)
- [Assessing Step-by-Step Reasoning against Lexical Negation: A Case Study on Syllogism](https://mlsanity.com/paper/2310.14868/) (2023)
- [Encoding Explanatory Knowledge for Zero-shot Science Question Answering](https://mlsanity.com/paper/2105.05737/) (2021)
- [Consecutive Question Generation via Dynamic Multitask Learning](https://mlsanity.com/paper/2211.08850/) (2022)
- [Lite Unified Modeling for Discriminative Reading Comprehension](https://mlsanity.com/paper/2203.14103/) (2022)
- [Multi-grained Evidence Inference for Multi-choice Reading Comprehension](https://mlsanity.com/paper/2310.18070/) (2023)
- [PerfMamba: Performance Analysis and Pruning of Selective State Space Models](https://mlsanity.com/paper/2511.22849/) (2025)
- [ActKnow: Active External Knowledge Infusion Learning for Question Answering in Low Data Regime](https://mlsanity.com/paper/2112.09423/) (2021)
- [Deep Understanding based Multi-Document Machine Reading Comprehension](https://mlsanity.com/paper/2204.03494/) (2022)
- [An Understanding-Oriented Robust Machine Reading Comprehension Model](https://mlsanity.com/paper/2207.00187/) (2022)
- [EMBRACE: Evaluation and Modifications for Boosting RACE](https://mlsanity.com/paper/2305.08433/) (2023)
- [Differentiating Choices via Commonality for Multiple-Choice Question Answering](https://mlsanity.com/paper/2408.11554/) (2024)

## Abstract (arXiv)

Reading strategies have been shown to improve comprehension levels, especially for readers lacking adequate prior knowledge. Just as the process of knowledge accumulation is time-consuming for human readers, it is resource-demanding to impart rich general domain knowledge into a deep language model via pre-training. Inspired by reading strategies identified in cognitive science, and given limited computational resources -- just a pre-trained model and a fixed number of training instances -- we propose three general strategies aimed to improve non-extractive machine reading comprehension (MRC): (i) BACK AND FORTH READING that considers both the original and reverse order of an input sequence, (ii) HIGHLIGHTING, which adds a trainable embedding to the text embedding of tokens that are relevant to the question and candidate answers, and (iii) SELF-ASSESSMENT that generates practice questions and candidate answers directly from the text in an unsupervised manner. By fine-tuning a pre-trained language model (Radford et al., 2018) with our proposed strategies on the largest general domain multiple-choice MRC dataset RACE, we obtain a 5.8% absolute increase in accuracy over the previous best result achieved by the same pre-trained model fine-tuned on RACE without the use of strategies. We further fine-tune the resulting model on a target MRC task, leading to an absolute improvement of 6.2% in average accuracy over previous state-of-the-art approaches on six representative non-extractive MRC datasets from different domains (i.e., ARC, OpenBookQA, MCTest, SemEval-2018 Task 11, ROCStories, and MultiRC). These results demonstrate the effectiveness of our proposed strategies and the versatility and general applicability of our fine-tuned models that incorporate these strategies. Core code is available at https://github.com/nlpdata/strategy/.