mlsanityData to 14 Oct 2026

Sparse DNNs with Improved Adversarial Robustness

arXiv 1810.09619 ↗First posted 23 Oct 2018cs.LG cs.CR cs.CVcited by 72

Not on the map. Why?

Abstract (arXiv)

Deep neural networks (DNNs) are computationally/memory-intensive and vulnerable to adversarial attacks, making them prohibitive in some real-world applications. By converting dense models into sparse ones, pruning appears to be a promising solution to reducing the computation/memory cost. This paper studies classification models, especially DNN-based ones, to demonstrate that there exists intrinsic relationships between their sparsity and adversarial robustness. Our analyses reveal, both theoretically and empirically, that nonlinear DNN-based classifiers behave differently under l2 attacks from some linear ones. We further demonstrate that an appropriately higher model sparsity implies better robustness of nonlinear DNNs, whereas over-sparsified models can be more difficult to resist adversarial examples.

Key facts. Cited by 72 papers in the map (data through 14 Oct 2026). First posted on arXiv 23 Oct 2018.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 12; 2022: 29; 2023: 8; 2024: 10; 2025: 8; 2026: 5.

72 papers.

Sort:

By area:

  1. 2022The Lazy Neuron Phenomenon: On Emergence of Activation Sparsity in Transformerscited by 122
  2. 2023Structured Pruning for Deep Convolutional Neural Networks: A surveycited by 122
  3. 2022The Unreasonable Effectiveness of Random Pruning: Return of the Most Naive Baseline for Sparse Trainingcited by 85
  4. 2021Lost in Pruning: The Effects of Pruning Neural Networks beyond Test Accuracycited by 62
  5. 2022Pruning has a disparate impact on model accuracycited by 44
  6. 2022Sparsity Winning Twice: Better Robust Generalization from More Efficient Trainingcited by 41
  7. 2021Drawing Robust Scratch Tickets: Subnetworks with Inborn Robustness Are Found within Randomly Initialized Networkscited by 29
  8. 2023Sparsity May Cry: Let Us Fail (Current) Sparse Neural Networks Together!cited by 27
  9. 2024Pruning for Protection: Increasing Jailbreak Resistance in Aligned LLMs Without Fine-Tuningcited by 26
  10. 2021CIFS: Improving Adversarial Robustness of CNNs via Channel-wise Importance-based Feature Selectioncited by 21
  11. 2022Masking Adversarial Damage: Finding Adversarial Saliency for Robust and Sparse Networkcited by 21
  12. 2022Sparsity and Heterogeneous Dropout for Continual Learning in the Null Space of Neural Activationscited by 19
38 more
  1. 2022Towards Sparsification of Graph Neural Networkscited by 18
  2. 2024Sparse maximal update parameterization: A holistic approach to sparse training dynamicscited by 16
  3. 2021Simon Says: Evaluating and Mitigating Bias in Pruned Neural Networks with Knowledge Distillationcited by 15
  4. 2022Quarantine: Sparsity Can Uncover the Trojan Attack Trigger for Freecited by 15
  5. 2023Temporal Coherent Test-Time Optimization for Robust Video Classificationcited by 15
  6. 2022Pruning Adversarially Robust Neural Networks without Adversarial Examplescited by 12
  7. 2022How Robust is Your Fairness? Evaluating and Sustaining Fairness under Unseen Distribution Shiftscited by 11
  8. 2024Adversarial Pruning: A Survey and Benchmark of Pruning Methods for Adversarial Robustnesscited by 10
  9. 2021Recent Advances in Large Margin Learningcited by 7
  10. 2023Relationship between Model Compression and Adversarial Robustness: A Review of Current Evidencecited by 6
  11. 2024Investigating Calibration and Corruption Robustness of Post-hoc Pruned Perception CNNs: An Image Classification Benchmark Studycited by 5
  12. 2021Explaining Adversarial Vulnerability with a Data Sparsity Hypothesiscited by 4
  13. 2021Pruning in the Face of Adversariescited by 4
  14. 2022CSTAR: Towards Compact and STructured Deep Neural Networks with Adversarial Robustnesscited by 4
  15. 2025Robust Experts: the Effect of Adversarial Training on CNNs with Sparse Mixture-of-Experts Layerscited by 4
  16. 2021The Impact of Activation Sparsity on Overfitting in Convolutional Neural Networkscited by 3
  17. 2022Visual Attention Emerges from Recurrent Sparse Reconstructioncited by 3
  18. 2022Wavelet Regularization Benefits Adversarial Trainingcited by 3
  19. 2022An Adversarial Robustness Perspective on the Topology of Neural Networkscited by 3
  20. 2023GRAN: Ghost Residual Attention Network for Single Image Super Resolutioncited by 3
  21. 2023iPINNs: Incremental learning for Physics-informed neural networkscited by 3
  22. 2024Navigating Extremes: Dynamic Sparsity in Large Output Spacescited by 3
  23. 2022Deadwooding: Robust Global Pruning for Deep Neural Networkscited by 2
  24. 2022Surprises in adversarially-trained linear regressioncited by 2
  25. 2022RIGID: Robust Linear Regression with Missing Datacited by 2
  26. 2022Biological connectomes as a representation for the architecture of artificial neural networkscited by 2
  27. 2022PathProx: A Proximal Gradient Algorithm for Weight Decay Regularized Deep Neural Networkscited by 2
  28. 2023Leveraging Low-Rank and Sparse Recurrent Connectivity for Robust Closed-Loop Controlcited by 2
  29. 2024Specification Overfitting in Artificial Intelligencecited by 2
  30. 2025On the Interaction of Compressibility and Adversarial Robustnesscited by 2
  31. 2025Sparsity-Driven Plasticity in Multi-Task Reinforcement Learningcited by 2
  32. 2026Visual Token Compression Enhances Robustness of MLLMscited by 2
  33. 2021Dual Head Adversarial Trainingcited by 1
  34. 2021Resilience from Diversity: Population-based approach to harden models against adversarial attackscited by 1
  35. 2022Sparse Super-Regular Networkscited by 1
  36. 2022Finding Dynamics Preserving Adversarial Winning Ticketscited by 1
  37. 2022On the Robustness and Anomaly Detection of Sparse Neural Networkscited by 1
  38. 2022State-driven Implicit Modeling for Sparsity and Robustness in Neural Networkscited by 1

50 most cited of 72 All 72 (JSON)