Avoiding Reasoning Shortcuts: Adversarial Evaluation, Training, and Model Development for Multi-Hop QA
Open accessThe DOI matches its Crossref record; 14 of the 14 references it cites were found in a registry or an index.
Abstract
Multi-hop question answering requires a model to connect multiple pieces of evidence scattered in a long context to answer the question.In this paper, we show that in the multihop HotpotQA (Yang et al., 2018) dataset, the examples often contain reasoning shortcuts through which models can directly locate the answer by word-matching the question with a sentence in the context.We demonstrate this issue by constructing adversarial documents that create contradicting answers to the shortcut but do not affect the validity of the original answer.The performance of strong baseline models drops significantly on our adversarial evaluation, indicating that they are indeed exploiting the shortcuts rather than performing multi-hop reasoning.After adversarial training, the baseline's performance improves but is still limited on the adversarial evaluation.Hence, we use a control unit that dynamically attends to the question at different reasoning hops to guide the model's multihop reasoning.We show that this 2-hop model trained on the regular data is more robust to the adversaries than the baseline model.After adversarial training, this 2-hop model not only achieves improvements over its counterpart trained on regular data, but also outperforms the adversarially-trained 1-hop baseline.We hope that these insights and initial improvements will motivate the development of new models that combine explicit compositional reasoning with adversarial training.1
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Hochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
SQuAD: 100,000+ Questions for Machine Comprehension of Text
Rajpurkar, P., Zhang, J., Lopyrev, K., & Liang, P. (2016) · Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d16-1264 · Cited by 6,208 · cited 4×
“Jia and Liang (2017) first applied adversarial evaluation to QA models on the SQuAD (Rajpurkar et al., 2016) dataset by generating a sentence that only resembles the question syntactically and appending it to the paragraph.”
BackgroundMethodInfluentialMS MARCO: A Human Generated MAchine Reading COmprehension Dataset
Bajaj, P., Campos, D., Craswell, N., Deng, L., Gao, J., Liu, X., Majumder, R., McNamara, A., Mitra, B., Nguyen, T., et al. (2016) · arXiv · doi:10.48550/arxiv.1611.09268 · Cited by 2,473
Know What You Don't Know: Unanswerable Questions for SQuAD
Rajpurkar, P., Jia, R., & Liang, P. (2018) · arXiv · doi:10.48550/arxiv.1806.03822 · Cited by 2,242
Know What You Don’t Know: Unanswerable Questions for SQuAD
Rajpurkar, P., Jia, R., & Liang, P. (2018) · Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) · doi:10.18653/v1/p18-2124 · Cited by 2,240
HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering
Yang, Z., Qi, P., Zhang, S., Bengio, Y., Cohen, W., Salakhutdinov, R., & Manning, C. D. (2018) · Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d18-1259 · Cited by 1,804 · cited 9×
“Previous works have shown that models with strong question-aware context representation (Seo et al., 2017; Xiong et al., 2017) can achieve super-human performance on single-hop QA tasks like SQuAD (Rajpurkar et al., 2016, 2018). as QAngaroo (Welbl et al., 2017) and Hot-potQA (Yang et al., 2018), have been proposed to…”
BackgroundMethodInfluentialTriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension
Joshi, M., Choi, E., Weld, D., & Zettlemoyer, L. (2017) · Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/p17-1147 · Cited by 1,690
TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension
Joshi, M., Choi, E., Weld, D. S., & Zettlemoyer, L. (2017) · arXiv · doi:10.48550/arxiv.1705.03551 · Cited by 1,650
Reading Wikipedia to answer open-domain questions
Chen, D., Fisch, A., Weston, J., & Bordes, A. (2017) · Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/p17-1171 · Cited by 1,546 · cited 2×
“Yang et al. (2018) then select the top-8 documents from Wikipedia with the shortest bigram TF-IDF (Chen et al., 2017) distance to the question as the distractors to form the context with a total of 10 documents.”
BackgroundMethod
Cited by
Red Teaming Language Models with Language Models
Perez, E., Huang, S., Song, F., Cai, T., Ring, R., Aslanides, J., Glaese, A., McAleese, N., & Irving, G. (2022) · Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2022.emnlp-main.225 · Cited by 363
Did Aristotle Use a Laptop? A Question Answering Benchmark with Implicit Reasoning Strategies
Geva, M., Khashabi, D., Segal, E., Khot, T., Roth, D., & Berant, J. (2021) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00370 · Cited by 242
“Moreover, answering a question based on relevant context does not imply that the model performs reasoning properly (Jiang and Bansal, 2019).”
BackgroundHierarchical Graph Network for Multi-hop Question Answering
Fang, Y., Sun, S., Gan, Z., Pillai, R., Wang, S., & Liu, J. (2020) · Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2020.emnlp-main.710 · Cited by 176
Knowledge Graph Prompting for Multi-Document Question Answering
Wang, Y., Lipka, N., Rossi, R. A., Siu, A., Zhang, R., & Derr, T. (2023) · arXiv · doi:10.48550/arxiv.2308.11730 · Cited by 170
Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned
Ganguli, D., Lovitt, L., Kernion, J., Askell, A., Bai, Y., Kadavath, S., Mann, B., Perez, E., Schiefer, N., Ndousse, K., et al. (2022) · arXiv · doi:10.48550/arxiv.2209.07858 · Cited by 130
“Some of these efforts augment humans (through guidelines, templates, programmatic generation of attacks, and various combinations thereof) to devise test cases that cause systems to fail [45, 46, 29, 21, 30, 55, 6, 23].”
Background
Metadata
- Authors
- Jiang, Y., & Bansal, M.
- Venue
- Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics
- Published
- 2019
- Pages
- 2726-2736
- OpenAlex
- W2951328433
- Cited by
- 83
- References
- 14
- FWCI
- 6.21
- Type
- conference paper
- Language
- en
- Open access
- hybrid
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Avoiding Reasoning Shortcuts: Adversarial Evaluation, Training, and Model Development for Multi-Hop QA. https://app.cochat.ai/papers/10.18653/v1/p19-1262
- The paper
- Jiang, Y., & Bansal, M. (2019). Avoiding Reasoning Shortcuts: Adversarial Evaluation, Training, and Model Development for Multi-Hop QA. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. https://doi.org/10.18653/v1/p19-1262
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.