mlsanityData to 7 Oct 2026

MergeGuard: Efficient Thwarting of Trojan Attacks in Machine Learning Models

arXiv 2505.04015 ↗First posted 6 May 2025cs.CR cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

This paper proposes MergeGuard, a novel methodology for mitigation of AI Trojan attacks. Trojan attacks on AI models cause inputs embedded with triggers to be misclassified to an adversary's target class, posing a significant threat to model usability trained by an untrusted third party. The core of MergeGuard is a new post-training methodology for linearizing and merging fully connected layers which we show simultaneously improves model generalizability and performance. Our Proof of Concept evaluation on Transformer models demonstrates that MergeGuard maintains model accuracy while decreasing trojan attack success rate, outperforming commonly used (post-training) Trojan mitigation by fine-tuning methodologies.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 20 of its 28 references that are in the map. First posted on arXiv 6 May 2025.

Builds on

20 of 28 references.

References · 20

  1. 2020Scaling Laws for Neural Language Modelscited by 6,411
  2. 2015Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classificationcited by 3,590
  3. 2016Gaussian Error Linear Units (GELUs)cited by 3,431
  4. 2015Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)cited by 1,061
  5. 2017BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chaincited by 1,052
  6. 2017Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoningcited by 1,012
  7. 2017Sigmoid-Weighted Linear Units for Neural Network Function Approximation in Reinforcement Learningcited by 935
  8. 2021AST: Audio Spectrogram Transformercited by 634
  9. 2018Fine-Pruning: Defending Against Backdooring Attacks on Deep Neural Networkscited by 550
  10. 2020Backdoor Learning: A Surveycited by 412
  11. 2021Anti-Backdoor Learning: Training Clean Models on Poisoned Datacited by 258
  12. 2019A new Backdoor Attack in CNNs by training set corruption without label poisoningcited by 237
8 more
  1. 2022ST-Adapter: Parameter-Efficient Image-to-Video Transfer Learningcited by 192
  2. 2017Neural Trojanscited by 151
  3. 2022BackdoorBench: A Comprehensive Benchmark of Backdoor Learningcited by 124
  4. 2021Model Complexity of Deep Learning: A Surveycited by 64
  5. 2023Enhancing Fine-Tuning Based Backdoor Defense with Sharpness-Aware Minimizationcited by 60
  6. 2020Computing Systems for Autonomous Driving: State-of-the-Art and Challengescited by 47
  7. 2020CLEANN: Accelerated Trojan Shield for Embedded Neural Networkscited by 22
  8. 2021Trojan Signatures in DNN Weightscited by 15

Built on by

Not cited yet.