mlsanityData to 7 Oct 2026

Towards Debiasing Fact Verification Models

arXiv 1908.05267 ↗First posted 14 Aug 2019cs.CLcited by 143

Not on the map. Why?

Abstract (arXiv)

Fact verification requires validating a claim in the context of evidence. We show, however, that in the popular FEVER dataset this might not necessarily be the case. Claim-only classifiers perform competitively with top evidence-aware models. In this paper, we investigate the cause of this phenomenon, identifying strong cues for predicting labels solely based on the claim, without considering any evidence. We create an evaluation set that avoids those idiosyncrasies. The performance of FEVER-trained models significantly drops when evaluated on this test set. Therefore, we introduce a regularization method which alleviates the effect of bias in the training data, obtaining improvements on the newly created test set. This work is a step towards a more sound evaluation of reasoning capabilities in fact verification models.

Key facts. Cited by 143 papers in the map (data through 7 Oct 2026). First posted on arXiv 14 Aug 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 37; 2022: 34; 2023: 26; 2024: 23; 2025: 13; 2026: 10.

143 papers.

Sort:

By area:

  1. 2021A Survey on Automated Fact-Checkingcited by 376
  2. 2021Get Your Vitamin C! Robust Fact Verification with Contrastive Evidencecited by 222
  3. 2021FEVEROUS: Fact Extraction and VERification Over Unstructured and Structured informationcited by 166
  4. 2021Gender Bias in Machine Translationcited by 116
  5. 2022Shortcut Learning of Large Language Models in Natural Language Understandingcited by 83
  6. 2021Towards Interpreting and Mitigating Shortcut Learning Behavior of NLU Modelscited by 78
  7. 2021Open-Domain, Content-based, Multi-modal Fact-checking of Out-of-Context Images via Online Resourcescited by 71
  8. 2021DialFact: A Benchmark for Fact-Checking in Dialoguecited by 62
  9. 2022Handling Bias in Toxic Speech Detection: A Surveycited by 60
  10. 2023FactKG: Fact Verification via Reasoning on Knowledge Graphscited by 57
  11. 2022Generating Data to Mitigate Spurious Correlations in Natural Language Inference Datasetscited by 54
  12. 2021ProoFVer: Natural Logic Theorem Proving for Fact Verificationcited by 52
38 more
  1. 2021Variational Information Bottleneck for Effective Low-Resource Fine-Tuningcited by 51
  2. 2023SCITAB: A Challenging Benchmark for Compositional Reasoning and Claim Verification on Scientific Tablescited by 51
  3. 2021Fool Me Twice: Entailment from Wikipedia Gamificationcited by 50
  4. 2023Do Large Language Models Know about Facts?cited by 49
  5. 2022Missing Counter-Evidence Renders NLP Fact-Checking Unrealistic for Misinformationcited by 45
  6. 2022Stretching Sentence-pair NLI Models to Reason over Long Documents and Clusterscited by 44
  7. 2021Zero-shot Fact Verification by Claim Generationcited by 42
  8. 2021AmbiFC: Fact-Checking Ambiguous Claims with Evidencecited by 38
  9. 2022Generalizing to the Future: Mitigating Entity Bias in Fake News Detectioncited by 38
  10. 2021FaVIQ: FAct Verification from Information-seeking Questionscited by 36
  11. 2023EX-FEVER: A Dataset for Multi-hop Explainable Fact Verificationcited by 36
  12. 2023Language Models Hallucinate, but May Excel at Fact Verificationcited by 35
  13. 2024Causal Prompting: Debiasing Large Language Model Prompting based on Front-Door Adjustmentcited by 29
  14. 2021Training Dynamics for Text Summarization Modelscited by 25
  15. 2023Check-COVID: Fact-Checking COVID-19 News Claims with Scientific Evidencecited by 25
  16. 2021End-to-End Self-Debiasing Framework for Robust NLU Trainingcited by 24
  17. 2022Deconfounding Legal Judgment Prediction for European Court of Human Rights Cases Towards Better Alignment with Expertscited by 24
  18. 2021Avoiding Inference Heuristics in Few-shot Prompt-based Finetuningcited by 23
  19. 2021CrossAug: A Contrastive Data Augmentation Method for Debiasing Fact Verification Modelscited by 23
  20. 2021Did they answer? Subjective acts and intents in conversational discoursecited by 22
  21. 2023A Survey on Out-of-Distribution Evaluation of Neural NLP Modelscited by 20
  22. 2021Context-aware Adversarial Training for Name Regularity Bias in Named Entity Recognitioncited by 19
  23. 2024Generative Large Language Models in Automated Fact-Checking: A Surveycited by 19
  24. 2021RAIL-KD: RAndom Intermediate Layer Mapping for Knowledge Distillationcited by 18
  25. 2021Debiasing Methods in Natural Language Understanding Make Bias More Accessiblecited by 17
  26. 2021Predict then Interpolate: A Simple Algorithm to Learn Stable Classifierscited by 16
  27. 2021Explainable Fact-checking through Question Answeringcited by 16
  28. 2021Uncertainty Calibration for Ensemble-Based Debiasing Methodscited by 16
  29. 2022Learning to Split for Automatic Bias Detectioncited by 16
  30. 2022Does Your Model Classify Entities Reasonably? Diagnosing and Mitigating Spurious Correlations in Entity Typingcited by 16
  31. 2022Debiasing Masks: A New Framework for Shortcut Mitigation in NLUcited by 16
  32. 2022Finding Dataset Shortcuts with Grammar Inductioncited by 15
  33. 2024Do We Need Language-Specific Fact-Checking Models? The Case of Chinesecited by 15
  34. 2022Less Learn Shortcut: Analyzing and Mitigating Learning of Spurious Feature-Label Correlationcited by 13
  35. 2021Robustness Challenges in Model Distillation and Pruning for Natural Language Understandingcited by 12
  36. 2021Dataset of Fake News Detection and Fact Verification: A Surveycited by 11
  37. 2022Fact Checking with Insufficient Evidencecited by 11
  38. 2022Are Prompt-based Models Clueless?cited by 11

50 most cited of 143 All 143 (JSON)