mlsanityData to 14 Oct 2026

Model Compression with Adversarial Robustness: A Unified Optimization Framework

arXiv 1902.03538 ↗First posted 10 Feb 2019cs.LG stat.MLcited by 71

Not on the map. Why?

Abstract (arXiv)

Deep model compression has been extensively studied, and state-of-the-art methods can now achieve high compression ratios with minimal accuracy loss. This paper studies model compression through a different lens: could we compress models without hurting their robustness to adversarial attacks, in addition to maintaining accuracy? Previous literature suggested that the goals of robustness and compactness might sometimes contradict. We propose a novel Adversarially Trained Model Compression (ATMC) framework. ATMC constructs a unified constrained optimization formulation, where existing compression means (pruning, factorization, quantization) are all integrated into the constraints. An efficient algorithm is then developed. An extensive group of experiments are presented, demonstrating that ATMC obtains remarkably more favorable trade-off among model size, accuracy and robustness, over currently available alternatives in various settings. The codes are publicly available at: https://github.com/shupenggui/ATMC.

Key facts. Cited by 71 papers in the map (data through 14 Oct 2026). First posted on arXiv 10 Feb 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 16; 2022: 25; 2023: 10; 2024: 13; 2025: 6; 2026: 1.

71 papers.

Sort:

By area:

  1. 2023Structured Pruning for Deep Convolutional Neural Networks: A surveycited by 122
  2. 2021AugMax: Adversarial Composition of Random Augmentations for Robust Trainingcited by 90
  3. 2022The Unreasonable Effectiveness of Random Pruning: Return of the Most Naive Baseline for Sparse Trainingcited by 85
  4. 2022A Survey on Long-Tailed Visual Recognitioncited by 75
  5. 2021Lost in Pruning: The Effects of Pruning Neural Networks beyond Test Accuracycited by 62
  6. 2022Sparsity Winning Twice: Better Robust Generalization from More Efficient Trainingcited by 41
  7. 2021Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safetycited by 35
  8. 2023Robust Mixture-of-Expert Training for Convolutional Neural Networkscited by 30
  9. 2021GDP: Stabilized Neural Network Pruning via Gates with Differentiable Polarizationcited by 29
  10. 2021Robust Predictable Controlcited by 29
  11. 2022Machine Learning for Microcontroller-Class Hardware: A Reviewcited by 28
  12. 2024Pruning for Protection: Increasing Jailbreak Resistance in Aligned LLMs Without Fine-Tuningcited by 26
38 more
  1. 2021Taxonomy of Machine Learning Safety: A Survey and Primercited by 25
  2. 2024Pruning for Robust Concept Erasing in Diffusion Modelscited by 22
  3. 2022Masking Adversarial Damage: Finding Adversarial Saliency for Robust and Sparse Networkcited by 21
  4. 2021Stealthy Backdoors as Compression Artifactscited by 18
  5. 2022Robust Lottery Tickets for Pre-trained Language Modelscited by 16
  6. 2022GhostNets on Heterogeneous Devices via Cheap Operationscited by 15
  7. 2022Quarantine: Sparsity Can Uncover the Trojan Attack Trigger for Freecited by 15
  8. 2022Can pruning improve certified robustness of neural networks?cited by 14
  9. 2023Robust low-rank training via approximate orthonormal constraintscited by 13
  10. 2025Adversarial Examples Are Not Bugs, They Are Superpositioncited by 13
  11. 2022Pruning Adversarially Robust Neural Networks without Adversarial Examplescited by 12
  12. 2021AdaSTE: An Adaptive Straight-Through Estimator to Train Binary Neural Networkscited by 11
  13. 2024Fast and Effective Weight Update for Pruned Large Language Modelscited by 11
  14. 2022Linearity Grafting: Relaxed Neuron Pruning Helps Certifiable Robustnesscited by 10
  15. 2022A Secure and Efficient Federated Learning Framework for NLPcited by 9
  16. 2021Learning Transferable 3D Adversarial Cloaks for Deep Trained Detectorscited by 8
  17. 2021Integer-arithmetic-only Certified Robustness for Quantized Neural Networkscited by 7
  18. 2022A Win-win Deal: Towards Sparse and Robust Pre-trained Language Modelscited by 7
  19. 2023SparseVSR: Lightweight and Noise Robust Visual Speech Recognitioncited by 7
  20. 2023Towards Robust Pruning: An Adaptive Knowledge-Retention Pruning Strategy for Language Modelscited by 7
  21. 2022A Tale of Two Models: Constructing Evasive Attacks on Edge Modelscited by 6
  22. 2023Relationship between Model Compression and Adversarial Robustness: A Review of Current Evidencecited by 6
  23. 20212-in-1 Accelerator: Enabling Random Precision Switch for Winning Both Adversarial Robustness and Efficiencycited by 5
  24. 2022Robust Binary Models by Pruning Randomly-initialized Networkscited by 5
  25. 2022Data-Efficient Double-Win Lottery Tickets from Robust Pre-trainingcited by 5
  26. 2024Investigating Calibration and Corruption Robustness of Post-hoc Pruned Perception CNNs: An Image Classification Benchmark Studycited by 5
  27. 2022CSTAR: Towards Compact and STructured Deep Neural Networks with Adversarial Robustnesscited by 4
  28. 2021Generalized Depthwise-Separable Convolutions for Adversarially Robust and Efficient Neural Networkscited by 3
  29. 2021Low-Rank+Sparse Tensor Compression for Neural Networkscited by 2
  30. 2022DASS: Differentiable Architecture Search for Sparse neural networkscited by 2
  31. 2023RNAS-CL: Robust Neural Architecture Search by Cross-Layer Knowledge Distillationcited by 2
  32. 2023Hint-Aug: Drawing Hints from Foundation Vision Transformers Towards Boosted Few-Shot Parameter-Efficient Tuningcited by 2
  33. 2024Transferable Learned Image Compression-Resistant Adversarial Perturbationscited by 2
  34. 2024Non-transferable Pruningcited by 2
  35. 2025On the Interaction of Compressibility and Adversarial Robustnesscited by 2
  36. 2022Finding Dynamics Preserving Adversarial Winning Ticketscited by 1
  37. 2022A Theoretical Study of The Effects of Adversarial Attacks on Sparse Regressioncited by 1
  38. 2023Samples on Thin Ice: Re-Evaluating Adversarial Pruning of Neural Networkscited by 1

50 most cited of 71 All 71 (JSON)