Evaluating Model Bias Requires Characterizing its Mistakes

• URL: http://arxiv.org/abs/2407.10633v1
• Date: Mon, 15 Jul 2024 11:46:21 GMT
• Title: Evaluating Model Bias Requires Characterizing its Mistakes
• Authors: Isabela Albuquerque, Jessica Schrouff, David Warde-Farley, Taylan Cemgil, Sven Gowal, Olivia Wiles,
• Abstract summary: We introduce SkewSize, a principled and flexible metric that captures bias from mistakes in a model's predictions. It can be used in multi-class settings or generalised to the open vocabulary setting of generative models. We demonstrate the utility of SkewSize in multiple settings including: standard vision models trained on synthetic data, vision models trained on ImageNet, and large scale vision-and-language models from the BLIP-2 family.
• Score: 19.777130236160712
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The ability to properly benchmark model performance in the face of spurious correlations is important to both build better predictors and increase confidence that models are operating as intended. We demonstrate that characterizing (as opposed to simply quantifying) model mistakes across subgroups is pivotal to properly reflect model biases, which are ignored by standard metrics such as worst-group accuracy or accuracy gap. Inspired by the hypothesis testing framework, we introduce SkewSize, a principled and flexible metric that captures bias from mistakes in a model's predictions. It can be used in multi-class settings or generalised to the open vocabulary setting of generative models. SkewSize is an aggregation of the effect size of the interaction between two categorical variables: the spurious variable representing the bias attribute and the model's prediction. We demonstrate the utility of SkewSize in multiple settings including: standard vision models trained on synthetic data, vision models trained on ImageNet, and large scale vision-and-language models from the BLIP-2 family. In each case, the proposed SkewSize is able to highlight biases not captured by other metrics, while also providing insights on the impact of recently proposed techniques, such as instruction tuning.

Related papers

• CONTESTS: a Framework for Consistency Testing of Span Probabilities in Language Models [16.436592723426305]
  It is unclear whether language models produce the same value for different ways of assigning joint probabilities to word spans. Our work introduces a novel framework, ConTestS, involving statistical tests to assess score consistency across interchangeable completion and conditioning orders.
  arXiv  Detail & Related papers  (2024-09-30T06:24:43Z)
• Predictive Churn with the Set of Good Models [64.05949860750235]
  We study the effect of conflicting predictions over the set of near-optimal machine learning models. We present theoretical results on the expected churn between models within the Rashomon set. We show how our approach can be used to better anticipate, reduce, and avoid churn in consumer-facing applications.
  arXiv  Detail & Related papers  (2024-02-12T16:15:25Z)
• Variation of Gender Biases in Visual Recognition Models Before and After Finetuning [29.55318393877906]
  We introduce a framework to measure how biases change before and after fine-tuning a large scale visual recognition model for a downstream task. We find that supervised models trained on datasets such as ImageNet-21k are more likely to retain their pretraining biases. We also find that models finetuned on larger scale datasets are more likely to introduce new biased associations.
  arXiv  Detail & Related papers  (2023-03-14T03:42:47Z)
• Debiasing Vision-Language Models via Biased Prompts [79.04467131711775]
  We propose a general approach for debiasing vision-language foundation models by projecting out biased directions in the text embedding. We show that debiasing only the text embedding with a calibrated projection matrix suffices to yield robust classifiers and fair generative models.
  arXiv  Detail & Related papers  (2023-01-31T20:09:33Z)
• On the Strong Correlation Between Model Invariance and Generalization [54.812786542023325]
  Generalization captures a model's ability to classify unseen data. Invariance measures consistency of model predictions on transformations of the data. From a dataset-centric view, we find a certain model's accuracy and invariance linearly correlated on different test sets.
  arXiv  Detail & Related papers  (2022-07-14T17:08:25Z)
• Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time [69.7693300927423]
  We show that averaging the weights of multiple models fine-tuned with different hyper parameter configurations improves accuracy and robustness. We show that the model soup approach extends to multiple image classification and natural language processing tasks.
  arXiv  Detail & Related papers  (2022-03-10T17:03:49Z)
• A Gating Model for Bias Calibration in Generalized Zero-shot Learning [18.32369721322249]
  Generalized zero-shot learning (GZSL) aims at training a model that can generalize to unseen class data by only using auxiliary information. One of the main challenges in GZSL is a biased model prediction toward seen classes caused by overfitting on only available seen class data during training. We propose a two-stream autoencoder-based gating model for GZSL.
  arXiv  Detail & Related papers  (2022-03-08T16:41:06Z)
• Training Experimentally Robust and Interpretable Binarized Regression Models Using Mixed-Integer Programming [3.179831861897336]
  We present a model-based approach to training robust and interpretable binarized regression models for multiclass classification tasks. Our MIP model balances the optimization of prediction margin and model size by using a weighted objective. We show the effectiveness of training robust and interpretable binarized regression models using MIP.
  arXiv  Detail & Related papers  (2021-12-01T11:53:08Z)
• Post-mortem on a deep learning contest: a Simpson's paradox and the complementary roles of scale metrics versus shape metrics [61.49826776409194]
  We analyze a corpus of models made publicly-available for a contest to predict the generalization accuracy of neural network (NN) models. We identify what amounts to a Simpson's paradox: where "scale" metrics perform well overall but perform poorly on sub partitions of the data. We present two novel shape metrics, one data-independent, and the other data-dependent, which can predict trends in the test accuracy of a series of NNs.
  arXiv  Detail & Related papers  (2021-06-01T19:19:49Z)
• Characterizing Fairness Over the Set of Good Models Under Selective Labels [69.64662540443162]
  We develop a framework for characterizing predictive fairness properties over the set of models that deliver similar overall performance. We provide tractable algorithms to compute the range of attainable group-level predictive disparities. We extend our framework to address the empirically relevant challenge of selectively labelled data.
  arXiv  Detail & Related papers  (2021-01-02T02:11:37Z)
• A Causal Lens for Peeking into Black Box Predictive Models: Predictive Model Interpretation via Causal Attribution [3.3758186776249928]
  We aim to address this problem in settings where the predictive model is a black box. We reduce the problem of interpreting a black box predictive model to that of estimating the causal effects of each of the model inputs on the model output. We show how the resulting causal attribution of responsibility for model output to the different model inputs can be used to interpret the predictive model and to explain its predictions.
  arXiv  Detail & Related papers  (2020-08-01T23:20:57Z)

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.