mlsanityData to 14 Oct 2026

MaskTune: Mitigating Spurious Correlations by Forcing to Explore

arXiv 2210.00055 ↗First posted 30 Sep 2022cs.LG cs.CVcited by 66

Not on the map. Why?

Abstract (arXiv)

A fundamental challenge of over-parameterized deep learning models is learning meaningful data representations that yield good performance on a downstream task without over-fitting spurious input features. This work proposes MaskTune, a masking strategy that prevents over-reliance on spurious (or a limited number of) features. MaskTune forces the trained model to explore new features during a single epoch finetuning by masking previously discovered features. MaskTune, unlike earlier approaches for mitigating shortcut learning, does not require any supervision, such as annotating spurious features or labels for subgroup samples in a dataset. Our empirical results on biased MNIST, CelebA, Waterbirds, and ImagenNet-9L datasets show that MaskTune is effective on tasks that often suffer from the existence of spurious correlations. Finally, we show that MaskTune outperforms or achieves similar performance to the competing methods when applied to the selective classification (classification with rejection option) task. Code for MaskTune is available at https://github.com/aliasgharkhani/Masktune.

Key facts. Cited by 66 papers in the map (data through 14 Oct 2026). Builds on 37 of its 67 references that are in the map. First posted on arXiv 30 Sep 2022.

Builds on

37 of 67 references.

References · 37

  1. 2016Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localizationcited by 6,733
  2. 2014Deep Learning Face Attributes in the Wildcited by 3,693
  3. 2019Invariant Risk Minimizationcited by 1,854
  4. 2015Show, Attend and Tell: Neural Image Caption Generation with Visual Attentioncited by 1,423
  5. 2019Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalizationcited by 1,247
  6. 2017A Closer Look at Memorization in Deep Networkscited by 942
  7. 2017Selective Classification for Deep Neural Networkscited by 852
  8. 2018Recognition in Terra Incognitacited by 658
  9. 2019Score-CAM: Score-Weighted Visual Explanations for Convolutional Neural Networkscited by 582
  10. 2021Just Train Twice: Improving Group Robustness without Training Group Informationcited by 523
  11. 2017Residual Attention Network for Image Classificationcited by 386
  12. 2020The Pitfalls of Simplicity Bias in Neural Networkscited by 351
25 more
  1. 2019SelectiveNet: A Deep Neural Network with an Integrated Reject Optioncited by 350
  2. 2020Noise or Signal: The Role of Image Backgrounds in Object Recognitioncited by 343
  3. 2020Gradient Starvation: A Learning Proclivity in Neural Networkscited by 267
  4. 2019Approximating CNNs with Bag-of-local-Features models works surprisingly well on ImageNetcited by 260
  5. 2017The Marginal Value of Adaptive Gradient Methods in Machine Learningcited by 250
  6. 2020No Subclass Left Behind: Fine-Grained Robustness in Coarse-Grained Classification Problemscited by 209
  7. 2019SGD on Neural Networks Learns Functions of Increasing Complexitycited by 204
  8. 2019Full-Gradient Representation for Neural Network Visualizationcited by 191
  9. 2016Statistics of Robust Optimization: A Generalized Empirical Likelihood Approachcited by 183
  10. 2018Deep learning generalizes because the parameter-function map is biased towards simple functionscited by 183
  11. 2020Large-Scale Methods for Distributionally Robust Optimizationcited by 179
  12. 2018Tell Me Where to Look: Guided Attention Inference Networkcited by 169
  13. 2019Adversarial Training Can Hurt Generalizationcited by 132
  14. 2021Counterfactual Generative Networkscited by 111
  15. 2019Deep Gamblers: Learning to Abstain with Portfolio Theorycited by 71
  16. 2020Self-training Avoids Using Spurious Features Under Domain Shiftcited by 69
  17. 2022Diversify and Disambiguate: Learning From Underspecified Datacited by 56
  18. 2020Removing Spurious Features can Hurt Accuracy and Affect Groups Disproportionatelycited by 52
  19. 2015Action Recognition using Visual Attentioncited by 49
  20. 2020Selective Classification via One-Sided Predictioncited by 38
  21. 2017Wasserstein Distributionally Robust Optimization and Variation Regularizationcited by 25
  22. 2021Robust Representation Learning via Perceptual Similarity Metricscited by 23
  23. 2020Learning Causal Models Onlinecited by 18
  24. 2013Intermittent process analysis with scattering momentscited by 10
  25. 2016Unanimous Prediction for 100% Precision with Application to Learning Semantic Mappingscited by 7

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 2; 2023: 12; 2024: 21; 2025: 19; 2026: 12.

66 papers.

Sort:

By area:

  1. 2023Towards Last-layer Retraining for Group Robustness with Fewer Annotationscited by 54
  2. 2024The Clever Hans Mirage: A Comprehensive Survey on Spurious Correlations in Machine Learningcited by 41
  3. 2025Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuningcited by 39
  4. 2023Towards Trustworthy and Aligned Machine Learning: A Data-centric Survey with Causality Perspectivescited by 30
  5. 2024Calibrating Multi-modal Representations: A Pursuit of Group Robustness without Annotationscited by 30
  6. 2023Group Robust Classification Without Any Group Informationcited by 25
  7. 2024Decompose-and-Compose: A Compositional Approach to Mitigating Spurious Correlationcited by 18
  8. 2023Studying How to Efficiently and Effectively Guide Models with Explanationscited by 16
  9. 2024B-cosification: Transforming Deep Neural Networks to be Inherently Interpretablecited by 12
  10. 2024Unsupervised Concept Discovery Mitigates Spurious Correlationscited by 11
  11. 2025Are We Done with Object-Centric Learning?cited by 11
  12. 2022Invariant Learning via Diffusion Dreamed Distribution Shiftscited by 10
38 more
  1. 2024Learning Robust Classifiers with Self-Guided Spurious Correlation Mitigationcited by 10
  2. 2024Sharpness-Aware Minimization Enhances Feature Quality via Balanced Learningcited by 9
  3. 2025Project-Probe-Aggregate: Efficient Fine-Tuning for Group Robustnesscited by 9
  4. 2022Towards Improved Input Masking for Convolutional Neural Networkscited by 8
  5. 2024Clarify: Improving Model Robustness With Natural Language Correctionscited by 8
  6. 2024ExMap: Leveraging Explainability Heatmaps for Unsupervised Group Robustness to Spurious Correlationscited by 8
  7. 2024Efficient Bias Mitigation Without Privileged Informationcited by 8
  8. 2025Mitigating Spurious Correlation via Distributionally Robust Learning with Hierarchical Ambiguity Setscited by 8
  9. 2023Mitigating Simplicity Bias in Deep Learning for Improved OOD Generalization and Robustnesscited by 7
  10. 2024Spuriousness-Aware Meta-Learning for Learning Robust Classifierscited by 7
  11. 2024Evaluating the Generalization Ability of Spatiotemporal Model in Urban Scenariocited by 7
  12. 2024The Group Robustness is in the Details: Revisiting Finetuning under Spurious Correlationscited by 6
  13. 2025Efficient Unsupervised Shortcut Learning Detection and Mitigation in Transformerscited by 6
  14. 2025DDB: Diffusion Driven Balancing to Address Spurious Correlationscited by 6
  15. 2023ASPIRE: Language-Guided Data Augmentation for Improving Robustness Against Spurious Correlationscited by 5
  16. 2024Towards Adversarial Robustness via Debiased High-Confidence Logit Alignmentcited by 5
  17. 2024Enhancing Intrinsic Features for Debiasing via Investigating Class-Discerning Common Attributes in Bias-Contrastive Paircited by 4
  18. 2025Gradient Extrapolation for Debiased Representation Learningcited by 4
  19. 2023Aerial Image Object Detection With Vision Transformer Detector (ViTDet)cited by 3
  20. 2023Debiased Automatic Speech Recognition for Dysarthric Speech via Sample Reweighting with Sample Affinity Testcited by 3
  21. 2023Partition-and-Debias: Agnostic Biases Mitigation via A Mixture of Biases-Specific Expertscited by 3
  22. 2023Disentangled Representation Learning with Transmitted Information Bottleneckcited by 3
  23. 2025BackMix: Regularizing Open Set Recognition by Removing Underlying Fore-Background Priorscited by 3
  24. 2025Attention IoU: Examining Biases in CelebA using Attention Mapscited by 3
  25. 2025On the Unreasonable Effectiveness of Last-layer Retrainingcited by 3
  26. 2023Remote Sensing Scene Classification with Masked Image Modeling (MIM)cited by 2
  27. 2024Learning in Order! A Sequential Strategy to Learn Invariant Features for Multimodal Sentiment Analysiscited by 2
  28. 2025LoRA Users Beware: A Few Spurious Tokens Can Manipulate Your Finetuned Modelcited by 2
  29. 2025AIM: Amending Inherent Interpretability via Self-Supervised Maskingcited by 2
  30. 2024Boosting Test Performance with Importance Sampling--a Subpopulation Perspectivecited by 1
  31. 2025Single-weight Model Editing for Post-hoc Spurious Correlation Neutralizationcited by 1
  32. 2025MiMu: Mitigating Multiple Shortcut Learning Behavior of Transformerscited by 1
  33. 2025Class-Conditional Distribution Balancing for Group Robust Classificationcited by 1
  34. 2025Fine-Grained Class-Conditional Distribution Balancing for Debiased Learningcited by 1
  35. 2025Two-stage Vision Transformers and Hard Masking offer Robust Object Representationscited by 1
  36. 2025Robust Context-Aware Object Recognitioncited by 1
  37. 2025Let Samples Speak: Mitigating Spurious Correlation by Exploiting the Clusterness of Samplescited by 1
  38. 2026HSFM: Hard-Set-Guided Feature-Space Meta-Learning for Robust Classification under Spurious Correlationscited by 1

50 most cited of 66 All 66 (JSON)