# MMM: Multi-stage Multi-task Learning for Multi-choice Reading Comprehension

arXiv: https://arxiv.org/abs/1910.00458 · HTML: https://mlsanity.com/paper/1910.00458/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 1 Oct 2019.

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

- [SciCap: Generating Captions for Scientific Figures](https://mlsanity.com/paper/2110.11624/) (2021)
- [Multi-Task Learning in Natural Language Processing: An Overview](https://mlsanity.com/paper/2109.09138/) (2021)
- [From LSAT: The Progress and Challenges of Complex Reasoning](https://mlsanity.com/paper/2108.00648/) (2021)
- [TSQA: Tabular Scenario Based Question Answering](https://mlsanity.com/paper/2101.11429/) (2021)
- [NeuralWOZ: Learning to Collect Task-Oriented Dialogue via Model-Based Simulation](https://mlsanity.com/paper/2105.14454/) (2021)
- [Multi-turn Dialogue Reading Comprehension with Pivot Turns and Knowledge](https://mlsanity.com/paper/2102.05474/) (2021)
- [ViMMRC 2.0 -- Enhancing Machine Reading Comprehension on Vietnamese Literature Text](https://mlsanity.com/paper/2303.18162/) (2023)
- [When Retriever-Reader Meets Scenario-Based Multiple-Choice Questions](https://mlsanity.com/paper/2108.13875/) (2021)
- [Towards Zero and Few-shot Knowledge-seeking Turn Detection in Task-orientated Dialogue Systems](https://mlsanity.com/paper/2109.08820/) (2021)
- [Learning-by-Narrating: Narrative Pre-Training for Zero-Shot Dialogue Comprehension](https://mlsanity.com/paper/2203.10249/) (2022)
- [Ensemble Learning-Based Approach for Improving Generalization Capability of Machine Reading Comprehension Systems](https://mlsanity.com/paper/2107.00368/) (2021)
- [Context-guided Triple Matching for Multiple Choice Question Answering](https://mlsanity.com/paper/2109.12996/) (2021)
- [XRJL-HKUST at SemEval-2021 Task 4: WordNet-Enhanced Dual Multi-head Co-Attention for Reading Comprehension of Abstract Meaning](https://mlsanity.com/paper/2103.16102/) (2021)
- [Semantic Sentence Composition Reasoning for Multi-Hop Question Answering](https://mlsanity.com/paper/2203.00160/) (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)
- [NEUer at SemEval-2021 Task 4: Complete Summary Representation by Filling Answers into Question for Matching Reading Comprehension](https://mlsanity.com/paper/2105.12051/) (2021)
- [A Question Answering Based Pipeline for Comprehensive Chinese EHR Information Extraction](https://mlsanity.com/paper/2402.11177/) (2024)
- [Transfer Learning Enhanced Single-choice Decision for Multi-choice Question Answering](https://mlsanity.com/paper/2404.17949/) (2024)
- [Investigating Recent Large Language Models for Vietnamese Machine Reading Comprehension](https://mlsanity.com/paper/2503.18062/) (2025)

## Abstract (arXiv)

Machine Reading Comprehension (MRC) for question answering (QA), which aims to answer a question given the relevant context passages, is an important way to test the ability of intelligence systems to understand human language. Multiple-Choice QA (MCQA) is one of the most difficult tasks in MRC because it often requires more advanced reading comprehension skills such as logical reasoning, summarization, and arithmetic operations, compared to the extractive counterpart where answers are usually spans of text within given passages. Moreover, most existing MCQA datasets are small in size, making the learning task even harder. We introduce MMM, a Multi-stage Multi-task learning framework for Multi-choice reading comprehension. Our method involves two sequential stages: coarse-tuning stage using out-of-domain datasets and multi-task learning stage using a larger in-domain dataset to help model generalize better with limited data. Furthermore, we propose a novel multi-step attention network (MAN) as the top-level classifier for this task. We demonstrate MMM significantly advances the state-of-the-art on four representative MCQA datasets.