mlsanityData to 7 Oct 2026

Generalist: Decoupling Natural and Robust Generalization

arXiv 2303.13813 ↗First posted 24 Mar 2023cs.CV cs.AI cs.LGcited by 14

Not on the map. Why?

Abstract (arXiv)

Deep neural networks obtained by standard training have been constantly plagued by adversarial examples. Although adversarial training demonstrates its capability to defend against adversarial examples, unfortunately, it leads to an inevitable drop in the natural generalization. To address the issue, we decouple the natural generalization and the robust generalization from joint training and formulate different training strategies for each one. Specifically, instead of minimizing a global loss on the expectation over these two generalization errors, we propose a bi-expert framework called Generalist where we simultaneously train base learners with task-aware strategies so that they can specialize in their own fields. The parameters of base learners are collected and combined to form a global learner at intervals during the training process. The global learner is then distributed to the base learners as initialized parameters for continued training. Theoretically, we prove that the risks of Generalist will get lower once the base learners are well trained. Extensive experiments verify the applicability of Generalist to achieve high accuracy on natural examples while maintaining considerable robustness to adversarial ones. Code is available at https://github.com/PKU-ML/Generalist.

Key facts. Cited by 14 papers in the map (data through 7 Oct 2026). Builds on 36 of its 46 references that are in the map. First posted on arXiv 24 Mar 2023.

Builds on

36 of 46 references.

References · 36

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  3. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  4. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  5. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  6. 2013Intriguing properties of neural networkscited by 5,114
  7. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,861
  8. 2016Wide Residual Networkscited by 3,057
  9. 2016Towards Evaluating the Robustness of Neural Networkscited by 2,976
  10. 2019Theoretically Principled Trade-off between Robustness and Accuracycited by 1,510
  11. 2020Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attackscited by 1,442
  12. 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,259
24 more
  1. 2018Averaging Weights Leads to Wider Optima and Better Generalizationcited by 1,194
  2. 2018On First-Order Meta-Learning Algorithmscited by 1,003
  3. 2018Robustness May Be at Odds with Accuracycited by 873
  4. 2015Distillation as a Defense to Adversarial Perturbations against Deep Neural Networkscited by 675
  5. 2017Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networkscited by 484
  6. 2019Unlabeled Data Improves Adversarial Robustnesscited by 415
  7. 2020Attacks Which Do Not Kill Training Make Adversarial Learning Strongercited by 242
  8. 2021Fixing Data Augmentation to Improve Adversarial Robustnesscited by 201
  9. 2019Are Labels Required for Improving Adversarial Robustness?cited by 162
  10. 2020Understanding and Mitigating the Tradeoff Between Robustness and Accuracycited by 148
  11. 2021On the Convergence and Robustness of Adversarial Trainingcited by 127
  12. 2016Fully-adaptive Feature Sharing in Multi-Task Networks with Applications in Person Attribute Classificationcited by 93
  13. 2020Adversarial Vertex Mixup: Toward Better Adversarially Robust Generalizationcited by 73
  14. 2016Deep Multi-task Representation Learning: A Tensor Factorisation Approachcited by 69
  15. 2022When Adversarial Training Meets Vision Transformers: Recipes from Training to Architecturecited by 64
  16. 2019Robustness to Adversarial Perturbations in Learning from Incomplete Datacited by 49
  17. 2019Interpolated Adversarial Training: Achieving Robust Neural Networks without Sacrificing Too Much Accuracycited by 49
  18. 2020Transferable, Controllable, and Inconspicuous Adversarial Attacks on Person Re-identification With Deep Mis-Rankingcited by 42
  19. 2022Self-Ensemble Adversarial Training for Improved Robustnesscited by 38
  20. 2019Detecting and Diagnosing Adversarial Images with Class-Conditional Capsule Reconstructionscited by 30
  21. 2015Fast and Accurate Recurrent Neural Network Acoustic Models for Speech Recognitioncited by 26
  22. 2016Integrated perception with recurrent multi-task neural networkscited by 24
  23. 2017Residual Convolutional CTC Networks for Automatic Speech Recognitioncited by 21
  24. 2020A Hamiltonian Monte Carlo Method for Probabilistic Adversarial Attack and Learningcited by 13

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 5; 2024: 3; 2025: 3; 2026: 3.

14 papers.

Sort:

By area:

  1. 2024Fight Back Against Jailbreaking via Prompt Adversarial Tuningcited by 63
  2. 2023CFA: Class-wise Calibrated Fair Adversarial Trainingcited by 53
  3. 2023Balance, Imbalance, and Rebalance: Understanding Robust Overfitting from a Minimax Game Perspectivecited by 18
  4. 2023Sharpness-Aware Minimization Alone can Improve Adversarial Robustnesscited by 14
  5. 2023MIMIR: Masked Image Modeling for Mutual Information-based Adversarial Robustnesscited by 12
  6. 2024Robustly Optimized Deep Feature Decoupling Network for Fatty Liver Diseases Detectioncited by 11
  7. 2026HoneyTrap: Deceiving Large Language Model Attackers to Honeypot Traps with Resilient Multi-Agent Defensecited by 6
  8. 2024FAIR-TAT: Improving Model Fairness Using Targeted Adversarial Trainingcited by 5
  9. 2023Splitting the Difference on Adversarial Trainingcited by 4
  10. 2025Generalist++: A Meta-learning Framework for Mitigating Trade-off in Adversarial Trainingcited by 1
  11. 2025Parameter Interpolation Adversarial Training for Robust Image Classificationcited by 1
  12. 2025Revisiting the Relationship between Adversarial and Clean Training: Why Clean Training Can Make Adversarial Training Bettercited by 0
2 more
  1. 2026Proxy Robustness in Vision Language Models is Effortlessly Transferablecited by 0
  2. 2026Unifying Adversarially Robust Model Experts in Vision-Language Modelscited by 0