# Reasoning Chain Based Adversarial Attack for Multi-hop Question Answering

arXiv: https://arxiv.org/abs/2112.09658 · HTML: https://mlsanity.com/paper/2112.09658/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 25 references that are in the map.
- First posted on arXiv 17 Dec 2021.

## Builds on (most cited first, 22 of 22)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [SQuAD: 100,000+ Questions for Machine Comprehension of Text](https://mlsanity.com/paper/1606.05250/) (2016)
- [HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering](https://mlsanity.com/paper/1809.09600/) (2018)
- [Know What You Don't Know: Unanswerable Questions for SQuAD](https://mlsanity.com/paper/1806.03822/) (2018)
- [Adversarial Examples for Evaluating Reading Comprehension Systems](https://mlsanity.com/paper/1707.07328/) (2017)
- [The Web as a Knowledge-base for Answering Complex Questions](https://mlsanity.com/paper/1803.06643/) (2018)
- [QASC: A Dataset for Question Answering via Sentence Composition](https://mlsanity.com/paper/1910.11473/) (2019)
- [Constructing Datasets for Multi-hop Reading Comprehension Across Documents](https://mlsanity.com/paper/1710.06481/) (2017)
- [Unsupervised Question Decomposition for Question Answering](https://mlsanity.com/paper/2002.09758/) (2020)
- [Hierarchical Graph Network for Multi-hop Question Answering](https://mlsanity.com/paper/1911.03631/) (2019)
- [Cognitive Graph for Multi-Hop Reading Comprehension at Scale](https://mlsanity.com/paper/1905.05460/) (2019)
- [Dynamically Fused Graph Network for Multi-hop Reasoning](https://mlsanity.com/paper/1905.06933/) (2019)
- [Avoiding Reasoning Shortcuts: Adversarial Evaluation, Training, and Model Development for Multi-Hop QA](https://mlsanity.com/paper/1906.07132/) (2019)
- [Select, Answer and Explain: Interpretable Multi-hop Reading Comprehension over Multiple Documents](https://mlsanity.com/paper/1911.00484/) (2019)
- [Question Answering by Reasoning Across Documents with Graph Convolutional Networks](https://mlsanity.com/paper/1808.09920/) (2018)
- [Multi-hop Reading Comprehension across Multiple Documents by Reasoning over Heterogeneous Graphs](https://mlsanity.com/paper/1905.07374/) (2019)
- [An Interpretable Reasoning Network for Multi-Relation Question Answering](https://mlsanity.com/paper/1801.04726/) (2018)
- [Robust Machine Comprehension Models via Adversarial Training](https://mlsanity.com/paper/1804.06473/) (2018)
- [Who did What: A Large-Scale Person-Centered Cloze Dataset](https://mlsanity.com/paper/1608.05457/) (2016)
- [Exploring Graph-structured Passage Representation for Multi-hop Reading Comprehension with Graph Neural Networks](https://mlsanity.com/paper/1809.02040/) (2018)
- [Multi-Paragraph Reasoning with Knowledge-enhanced Graph Neural Network](https://mlsanity.com/paper/1911.02170/) (2019)
- [Is Graph Structure Necessary for Multi-hop Question Answering?](https://mlsanity.com/paper/2004.03096/) (2020)

## Built on by (most cited first, 4 of 4)

- [Not All LLM Reasoners Are Created Equal](https://mlsanity.com/paper/2410.01748/) (2024)
- [Seemingly Plausible Distractors in Multi-Hop Reasoning: Are Large Language Models Attentive Readers?](https://mlsanity.com/paper/2409.05197/) (2024)
- [Teaching Broad Reasoning Skills for Multi-Step QA by Generating Hard Contexts](https://mlsanity.com/paper/2205.12496/) (2022)
- [CRiT-QA: Evaluating Multi-hop Reasoning with Counterfactual Chains and Distractor Traps](https://mlsanity.com/paper/2607.10562/) (2026)

## Abstract (arXiv)

Recent years have witnessed impressive advances in challenging multi-hop QA tasks. However, these QA models may fail when faced with some disturbance in the input text and their interpretability for conducting multi-hop reasoning remains uncertain. Previous adversarial attack works usually edit the whole question sentence, which has limited effect on testing the entity-based multi-hop inference ability. In this paper, we propose a multi-hop reasoning chain based adversarial attack method. We formulate the multi-hop reasoning chains starting from the query entity to the answer entity in the constructed graph, which allows us to align the question to each reasoning hop and thus attack any hop. We categorize the questions into different reasoning types and adversarially modify part of the question corresponding to the selected reasoning hop to generate the distracting sentence. We test our adversarial scheme on three QA models on HotpotQA dataset. The results demonstrate significant performance reduction on both answer and supporting facts prediction, verifying the effectiveness of our reasoning chain based attack method for multi-hop reasoning models and the vulnerability of them. Our adversarial re-training further improves the performance and robustness of these models.