mlsanityData to 14 Oct 2026

Patch Synthesis for Property Repair of Deep Neural Networks

arXiv 2404.01642 ↗First posted 2 Apr 2024cs.LG cs.CRcited by 0

Not on the map. Why?

Abstract (arXiv)

Deep neural networks (DNNs) are prone to various dependability issues, such as adversarial attacks, which hinder their adoption in safety-critical domains. Recently, NN repair techniques have been proposed to address these issues while preserving original performance by locating and modifying guilty neurons and their parameters. However, existing repair approaches are often limited to specific data sets and do not provide theoretical guarantees for the effectiveness of the repairs. To address these limitations, we introduce PatchPro, a novel patch-based approach for property-level repair of DNNs, focusing on local robustness. The key idea behind PatchPro is to construct patch modules that, when integrated with the original network, provide specialized repairs for all samples within the robustness neighborhood while maintaining the network's original performance. Our method incorporates formal verification and a heuristic mechanism for allocating patch modules, enabling it to defend against adversarial attacks and generalize to other inputs. PatchPro demonstrates superior efficiency, scalability, and repair success rates compared to existing DNN repair methods, i.e., realizing provable property-level repair for 100% cases across multiple high-dimensional datasets.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 45 of its 75 references that are in the map. First posted on arXiv 2 Apr 2024.

Builds on

45 of 75 references.

References · 45

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,364
  3. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  4. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
  5. 2016Wide Residual Networkscited by 3,070
  6. 2015Domain-Adversarial Training of Neural Networkscited by 3,044
  7. 2019Theoretically Principled Trade-off between Robustness and Accuracycited by 1,517
  8. 2016End to End Learning for Self-Driving Carscited by 1,009
  9. 2017Reluplex: An Efficient SMT Solver for Verifying Deep Neural Networkscited by 744
  10. 2018Efficient Neural Network Robustness Certification with General Activation Functionscited by 445
  11. 2016Safety Verification of Deep Neural Networkscited by 263
  12. 2018Towards Fast Computation of Certified Robustness for ReLU Networkscited by 255
33 more
  1. 2017DeepTest: Automated Testing of Deep-Neural-Network-driven Autonomous Carscited by 251
  2. 2017Formal Verification of Piece-Wise Linear Feed-Forward Neural Networkscited by 242
  3. 2019Adversarial Training and Robustness for Multiple Perturbationscited by 183
  4. 2020Feature Purification: How Adversarial Training Performs Robust Deep Learningcited by 107
  5. 2018Reachability Analysis of Deep Neural Networks with Provable Guaranteescited by 93
  6. 2020Verification of Deep Convolutional Neural Networks Using ImageStarscited by 75
  7. 2019An Abstraction-Based Framework for Neural Network Verificationcited by 68
  8. 2019Quantitative Verification of Neural Networks And its Security Applicationscited by 56
  9. 2018A Statistical Approach to Assessing Neural Network Robustnesscited by 54
  10. 2022Certified Training: Small Boxes are All You Needcited by 52
  11. 2021Mind the box: l1-APGD for sparse adversarial attacks on image classifierscited by 51
  12. 2017Feature-Guided Black-Box Safety Testing of Deep Neural Networkscited by 49
  13. 2018Adversarial Sample Detection for Deep Neural Network through Model Mutation Testingcited by 48
  14. 2022Causality-based Neural Network Repaircited by 47
  15. 2020Probabilistic Safety for Bayesian Neural Networkscited by 42
  16. 2021Provable Repair of Deep Neural Networkscited by 42
  17. 2019Robustness of Neural Networks: A Probabilistic and Practical Approachcited by 38
  18. 2019Statistical Guarantees for the Robustness of Bayesian Neural Networkscited by 35
  19. 2018A Game-Based Approximate Verification of Deep Neural Networks with Provable Guaranteescited by 33
  20. 2020Scalable Quantitative Verification For Deep Neural Networkscited by 31
  21. 2020DeepAbstract: Neural Network Abstraction for Accelerating Verificationcited by 31
  22. 2020ReluDiff: Differential Verification of Deep Neural Networkscited by 29
  23. 2019Analyzing Deep Neural Networks with Symbolic Propagation: Towards Higher Precision and Faster Verificationcited by 28
  24. 2019Arachne: Search Based Repair of Deep Neural Networkscited by 26
  25. 2022An Abstraction-Refinement Approach to Verifying Convolutional Neural Networkscited by 24
  26. 2018Robustness Guarantees for Bayesian Inference with Gaussian Processescited by 23
  27. 2017Residual Convolutional CTC Networks for Automatic Speech Recognitioncited by 21
  28. 2019ART: Abstraction Refinement-Guided Training for Provably Correct Neural Networkscited by 14
  29. 2021Sound and Complete Neural Network Repair with Minimality and Locality Guaranteescited by 13
  30. 2021Towards Practical Robustness Analysis for DNNs based on PAC-Model Learningcited by 12
  31. 2023Architecture-Preserving Provable Repair of Deep Neural Networkscited by 11
  32. 2019Verifying Aircraft Collision Avoidance Neural Networks Through Linear Approximations of Safe Regionscited by 9
  33. 2021ε-weakened Robustness of Deep Neural Networkscited by 8

Built on by

Not cited yet.