Corpus ID: 211076210

Adversarial Filters of Dataset Biases

@inproceedings{Bras2020AdversarialFO,
  title={Adversarial Filters of Dataset Biases},
  author={Ronan Le Bras and Swabha Swayamdipta and Chandra Bhagavatula and Rowan Zellers and Matthew E. Peters and A. Sabharwal and Yejin Choi},
  booktitle={ICML},
  year={2020}
}
  • Ronan Le Bras, Swabha Swayamdipta, +4 authors Yejin Choi
  • Published in ICML 2020
  • Computer Science, Mathematics
  • Large neural models have demonstrated human-level performance on language and vision benchmarks, while their performance degrades considerably on adversarial or out-of-distribution samples. This raises the question of whether these models have learned to solve a dataset rather than the underlying task by overfitting to spurious dataset biases. We investigate one recently proposed approach, AFLite, which adversarially filters such dataset biases, as a means to mitigate the prevalent… CONTINUE READING
    33 Citations
    Learning to Model and Ignore Dataset Bias with Mixed Capacity Ensembles
    • PDF
    Latent Adversarial Debiasing: Mitigating Collider Bias in Deep Neural Networks
    • Highly Influenced
    • PDF
    DQI: Measuring Data Quality in NLP
    • 5
    • Highly Influenced
    • PDF
    Improving QA Generalization by Concurrent Modeling of Multiple Biases
    • 4
    • PDF
    Generative Data Augmentation for Commonsense Reasoning
    • 2
    • PDF
    G-DAUG: Generative Data Augmentation for Commonsense Reasoning
    • 10
    • PDF
    Winning Ticket in Noisy Image Classification
    • PDF

    References

    SHOWING 1-10 OF 65 REFERENCES
    Inoculation by Fine-Tuning: A Method for Analyzing Challenge Datasets
    • 50
    • Highly Influential
    • PDF
    Natural Adversarial Examples
    • 94
    • PDF
    An Adversarial Winograd Schema Challenge at Scale
    • 1
    • PDF
    SWAG: A Large-Scale Adversarial Dataset for Grounded Commonsense Inference
    • 274
    • PDF
    Strike (With) a Pose: Neural Networks Are Easily Fooled by Strange Poses of Familiar Objects
    • 86
    • PDF
    REPAIR: Removing Representation Bias by Dataset Resampling
    • Y. Li, N. Vasconcelos
    • Computer Science
    • 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
    • 2019
    • 43
    • PDF
    Adversarial Removal of Demographic Attributes from Text Data
    • 101
    • PDF