mlsanityData to 7 Oct 2026

Dynamically Refined Regularization for Improving Cross-corpora Hate Speech Detection

arXiv 2203.12536 ↗First posted 23 Mar 2022cs.CLcited by 3

Not on the map. Why?

Abstract (arXiv)

Hate speech classifiers exhibit substantial performance degradation when evaluated on datasets different from the source. This is due to learning spurious correlations between words that are not necessarily relevant to hateful language, and hate speech labels from the training corpus. Previous work has attempted to mitigate this problem by regularizing specific terms from pre-defined static dictionaries. While this has been demonstrated to improve the generalizability of classifiers, the coverage of such methods is limited and the dictionaries require regular manual updates from human experts. In this paper, we propose to automatically identify and reduce spurious correlations using attribution methods with dynamic refinement of the list of terms that need to be regularized during training. Our approach is flexible and improves the cross-corpora performance over previous work independently and in combination with pre-defined dictionaries.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 23 of its 44 references that are in the map. First posted on arXiv 23 Mar 2022.

Builds on

23 of 44 references.

References · 23

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2017A Unified Approach to Interpreting Model Predictionscited by 6,488
  3. 2017Axiomatic Attribution for Deep Networkscited by 3,350
  4. 2017Learning Important Features Through Propagating Activation Differencescited by 1,454
  5. 2019Is Attention Interpretable?cited by 387
  6. 2017Right for the Right Reasons: Training Differentiable Models by Constraining their Explanationscited by 293
  7. 2020Learning from the Worst: Dynamically Generated Datasets to Improve Online Hate Detectioncited by 217
  8. 2017Deep Learning for Hate Speech Detection in Tweetscited by 180
  9. 2020Debugging Tests for Model Explanationscited by 146
  10. 2019Interpretations are useful: penalizing explanations to align neural networks with prior knowledgecited by 144
  11. 2019A BERT-Based Transfer Learning Approach for Hate Speech Detection in Online Social Mediacited by 95
  12. 2020Contextualizing Hate Speech Classifiers with Post-hoc Explanationcited by 84
11 more
  1. 2019Incorporating Priors with Feature Attribution on Text Classificationcited by 68
  2. 2021Towards generalisable hate speech detection: a review on obstacles and solutionscited by 66
  3. 2019Adapt or Get Left Behind: Domain Adaptation through BERT Language Model Finetuning for Aspect-Target Sentiment Classificationcited by 62
  4. 2020PERL: Pivot-based Domain Adaptation for Pre-trained Deep Contextualized Embedding Modelscited by 33
  5. 2021Refining Language Models with Compositional Explanationscited by 27
  6. 2019Towards Hierarchical Importance Attribution: Explaining Compositional Semantics for Neural Sequence Modelscited by 26
  7. 2022An Empirical Study on Explanations in Out-of-Domain Settingscited by 21
  8. 2020Knowledge Distillation for BERT Unsupervised Domain Adaptationcited by 14
  9. 2019Towards non-toxic landscapes: Automatic toxic comment detection using DNNcited by 8
  10. 2020Adversarial Augmentation Policy Search for Domain and Cross-Lingual Generalization in Reading Comprehensioncited by 7
  11. 2021Unsupervised Domain Adaptation for Hate Speech Detection Using a Data Augmentation Approachcited by 6

Built on by

Cited by year

Citing papers per year: 2022: 2; 2023: 1.

3 papers.

  1. 2023Robust Hate Speech Detection in Social Media: A Cross-Dataset Empirical Evaluationcited by 28
  2. 2022Calibration Meets Explanation: A Simple and Effective Approach for Model Confidence Estimatescited by 7
  3. 2022Domain Classification-based Source-specific Term Penalization for Domain Adaptation in Hate-speech Detectioncited by 0