mlsanityData to 14 Oct 2026

Confidence Matters: Inspecting Backdoors in Deep Neural Networks via Distribution Transfer

arXiv 2208.06592 ↗First posted 13 Aug 2022cs.CR cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Backdoor attacks have been shown to be a serious security threat against deep learning models, and detecting whether a given model has been backdoored becomes a crucial task. Existing defenses are mainly built upon the observation that the backdoor trigger is usually of small size or affects the activation of only a few neurons. However, the above observations are violated in many cases especially for advanced backdoor attacks, hindering the performance and applicability of the existing defenses. In this paper, we propose a backdoor defense DTInspector built upon a new observation. That is, an effective backdoor attack usually requires high prediction confidence on the poisoned training samples, so as to ensure that the trained model exhibits the targeted behavior with a high probability. Based on this observation, DTInspector first learns a patch that could change the predictions of most high-confidence data, and then decides the existence of backdoor by checking the ratio of prediction changes after applying the learned patch on the low-confidence data. Extensive evaluations on five backdoor attacks, four datasets, and three advanced attacking types demonstrate the effectiveness of the proposed defense.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 32 of its 54 references that are in the map. First posted on arXiv 13 Aug 2022.

Builds on

32 of 54 references.

References · 32

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2020Shortcut Learning in Deep Neural Networkscited by 1,969
  3. 2019Certified Adversarial Robustness via Randomized Smoothingcited by 1,340
  4. 2017BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chaincited by 1,106
  5. 2017Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoningcited by 1,055
  6. 2016Universal adversarial perturbationscited by 784
  7. 2018Fine-Pruning: Defending Against Backdooring Attacks on Deep Neural Networkscited by 580
  8. 2018Spectral Signatures in Backdoor Attackscited by 484
  9. 2019STRIP: A Defence Against Trojan Attacks on Deep Neural Networkscited by 459
  10. 2018Detecting Backdoor Attacks on Deep Neural Networks by Activation Clusteringcited by 456
  11. 2019Hidden Trigger Backdoor Attackscited by 371
  12. 2020Reflection Backdoor: A Natural Backdoor Attack on Deep Neural Networkscited by 349
20 more
  1. 2021Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networkscited by 324
  2. 2020Input-Aware Dynamic Backdoor Attackcited by 313
  3. 2021Anti-Backdoor Learning: Training Clean Models on Poisoned Datacited by 263
  4. 2019A new Backdoor Attack in CNNs by training set corruption without label poisoningcited by 242
  5. 2021Adversarial Neuron Pruning Purifies Backdoored Deep Modelscited by 235
  6. 2019Detecting AI Trojans Using Meta Neural Analysiscited by 199
  7. 2020Clean-Label Backdoor Attacks on Video Recognition Modelscited by 169
  8. 2020Dynamic Backdoor Attacks Against Machine Learning Modelscited by 151
  9. 2019Universal Litmus Patterns: Revealing Backdoor Attacks in CNNscited by 146
  10. 2016A vector-contraction inequality for Rademacher complexitiescited by 142
  11. 2020Backdoor Attacks Against Deep Learning Systems in the Physical Worldcited by 134
  12. 2020Bridging Mode Connectivity in Loss Landscapes and Adversarial Robustnesscited by 132
  13. 2021SPECTRE: Defending Against Backdoor Attacks Using Robust Statisticscited by 130
  14. 2020Backdoor Attacks and Countermeasures on Deep Learning: A Comprehensive Reviewcited by 123
  15. 2021Backdoor Scanning for Deep Neural Networks through K-Arm Optimizationcited by 98
  16. 2020An Embarrassingly Simple Approach for Trojan Attack in Deep Neural Networkscited by 97
  17. 2019Bypassing Backdoor Detection Algorithms in Deep Learningcited by 74
  18. 2019Model Agnostic Defence against Backdoor Attacks in Machine Learningcited by 52
  19. 2021Few-shot Backdoor Defense Using Shapley Estimationcited by 41
  20. 2020Live Trojan Attacks on Deep Neural Networkscited by 10

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2022BackdoorBench: A Comprehensive Benchmark of Backdoor Learning
  2. 2023BaDExpert: Extracting Backdoor Functionality for Accurate Backdoor Input Detection
  3. 2023PatchBackdoor: Backdoor Attack against Deep Neural Networks without Model Modification
  4. 2024BAN: Detecting Backdoors Activated by Adversarial Neuron Noise
  5. 2024Breaking the False Sense of Security in Backdoor Defense through Re-Activation Attack
  6. 2025DeBackdoor: A Deductive Framework for Detecting Backdoor Attacks on Deep Models with Limited Data
  7. 2023Beating Backdoor Attack at Its Own Game
  8. 2024Defending Deep Regression Models against Backdoor Attacks