Kernel-Whitening: Overcome Dataset Bias with Isotropic Sentence Embedding

• URL: http://arxiv.org/abs/2210.07547v1
• Date: Fri, 14 Oct 2022 05:56:38 GMT
• Title: Kernel-Whitening: Overcome Dataset Bias with Isotropic Sentence Embedding
• Authors: Songyang Gao, Shihan Dou, Qi Zhang, Xuanjing Huang
• Abstract summary: We propose a representation normalization method which aims at disentangling the correlations between features of encoded sentences. We also propose Kernel-Whitening, a Nystrom kernel approximation method to achieve more thorough debiasing on nonlinear spurious correlations. Experiments show that Kernel-Whitening significantly improves the performance of BERT on out-of-distribution datasets while maintaining in-distribution accuracy.
• Score: 51.48582649050054
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Dataset bias has attracted increasing attention recently for its detrimental effect on the generalization ability of fine-tuned models. The current mainstream solution is designing an additional shallow model to pre-identify biased instances. However, such two-stage methods scale up the computational complexity of training process and obstruct valid feature information while mitigating bias. To address this issue, we utilize the representation normalization method which aims at disentangling the correlations between features of encoded sentences. We find it also promising in eliminating the bias problem by providing isotropic data distribution. We further propose Kernel-Whitening, a Nystrom kernel approximation method to achieve more thorough debiasing on nonlinear spurious correlations. Our framework is end-to-end with similar time consumption to fine-tuning. Experiments show that Kernel-Whitening significantly improves the performance of BERT on out-of-distribution datasets while maintaining in-distribution accuracy.

Related papers

• Model Debiasing by Learnable Data Augmentation [19.625915578646758]
  This paper proposes a novel 2-stage learning pipeline featuring a data augmentation strategy able to regularize the training. Experiments on synthetic and realistic biased datasets show state-of-the-art classification accuracy, outperforming competing methods.
  arXiv  Detail & Related papers  (2024-08-09T09:19:59Z)
• Looking at Model Debiasing through the Lens of Anomaly Detection [11.113718994341733]
  Deep neural networks are sensitive to bias in the data. We propose a new bias identification method based on anomaly detection. We reach state-of-the-art performance on synthetic and real benchmark datasets.
  arXiv  Detail & Related papers  (2024-07-24T17:30:21Z)
• DiffInject: Revisiting Debias via Synthetic Data Generation using Diffusion-based Style Injection [9.801159950963306]
  We propose DiffInject, a powerful method to augment synthetic bias-conflict samples using a pretrained diffusion model. Our framework does not require any explicit knowledge of the bias types or labelling, making it a fully unsupervised setting for debiasing.
  arXiv  Detail & Related papers  (2024-06-10T09:45:38Z)
• Training Unbiased Diffusion Models From Biased Dataset [18.09610829650175]
  This paper proposes time-dependent importance reweighting to mitigate the bias for diffusion models. We demonstrate that the time-dependent density ratio becomes more precise than previous approaches. While directly applying it to score-matching is intractable, we discover that using the time-dependent density ratio both for reweighting and score correction can lead to a tractable form of the objective function.
  arXiv  Detail & Related papers  (2024-03-02T12:06:42Z)
• Revisiting the Dataset Bias Problem from a Statistical Perspective [72.94990819287551]
  We study the "dataset bias" problem from a statistical standpoint. We identify the main cause of the problem as the strong correlation between a class attribute u and a non-class attribute b. We propose to mitigate dataset bias via either weighting the objective of each sample n by frac1p(u_n|b_n) or sampling that sample with a weight proportional to frac1p(u_n|b_n).
  arXiv  Detail & Related papers  (2024-02-05T22:58:06Z)
• Self-supervised debiasing using low rank regularization [59.84695042540525]
  Spurious correlations can cause strong biases in deep neural networks, impairing generalization ability. We propose a self-supervised debiasing framework potentially compatible with unlabeled samples. Remarkably, the proposed debiasing framework significantly improves the generalization performance of self-supervised learning baselines.
  arXiv  Detail & Related papers  (2022-10-11T08:26:19Z)
• Predicting traffic signals on transportation networks using spatio-temporal correlations on graphs [56.48498624951417]
  This paper proposes a traffic propagation model that merges multiple heat diffusion kernels into a data-driven prediction model to forecast traffic signals. We optimize the model parameters using Bayesian inference to minimize the prediction errors and, consequently, determine the mixing ratio of the two approaches. The proposed model demonstrates prediction accuracy comparable to that of the state-of-the-art deep neural networks with lower computational effort.
  arXiv  Detail & Related papers  (2021-04-27T18:17:42Z)
• Decorrelated Clustering with Data Selection Bias [55.91842043124102]
  We propose a novel Decorrelation regularized K-Means algorithm (DCKM) for clustering with data selection bias. Our DCKM algorithm achieves significant performance gains, indicating the necessity of removing unexpected feature correlations induced by selection bias.
  arXiv  Detail & Related papers  (2020-06-29T08:55:50Z)
• Towards Robustifying NLI Models Against Lexical Dataset Biases [94.79704960296108]
  This paper explores both data-level and model-level debiasing methods to robustify models against lexical dataset biases. First, we debias the dataset through data augmentation and enhancement, but show that the model bias cannot be fully removed via this method. The second approach employs a bag-of-words sub-model to capture the features that are likely to exploit the bias and prevents the original model from learning these biased features.
  arXiv  Detail & Related papers  (2020-05-10T17:56:10Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.