mlsanityData to 13 Oct 2026

Dim-Krum: Backdoor-Resistant Federated Learning for NLP with Dimension-wise Krum-Based Aggregation

arXiv 2210.06894 ↗First posted 13 Oct 2022cs.LG cs.CRcited by 8

On the map: LLM safety and security → Federated unlearning and poisoning

Abstract (arXiv)

Despite the potential of federated learning, it is known to be vulnerable to backdoor attacks. Many robust federated aggregation methods are proposed to reduce the potential backdoor risk. However, they are mainly validated in the CV field. In this paper, we find that NLP backdoors are hard to defend against than CV, and we provide a theoretical analysis that the malicious update detection error probabilities are determined by the relative backdoor strengths. NLP attacks tend to have small relative backdoor strengths, which may result in the failure of robust federated aggregation methods for NLP attacks. Inspired by the theoretical results, we can choose some dimensions with higher backdoor strengths to settle this issue. We propose a novel federated aggregation algorithm, Dim-Krum, for NLP tasks, and experimental results validate its effectiveness.

Key facts. Cited by 8 papers in the map (data through 13 Oct 2026). Placed in Federated unlearning and poisoning, one of 12 directions in LLM safety and security. Builds on 33 of its 57 references that are in the map. First posted on arXiv 13 Oct 2022.

Builds on

33 of 57 references.

Closest ancestors · 1

Same direction, newest first

  1. 2021Robust Federated Learning with Attack-Adaptive Aggregation10 Feb

Other references · 32

  1. 2016Communication-Efficient Learning of Deep Networks from Decentralized Datacited by 7,653
  2. 2015Character-level Convolutional Networks for Text Classificationcited by 2,319
  3. 2014Convolutional Neural Networks for Sentence Classificationcited by 1,126
  4. 2017Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoningcited by 1,047
  5. 2018Byzantine-Robust Distributed Learning: Towards Optimal Statistical Ratescited by 708
  6. 2018Fine-Pruning: Defending Against Backdooring Attacks on Deep Neural Networkscited by 579
  7. 2019STRIP: A Defence Against Trojan Attacks on Deep Neural Networkscited by 456
  8. 2018Detecting Backdoor Attacks on Deep Neural Networks by Activation Clusteringcited by 451
  9. 2019Hidden Trigger Backdoor Attackscited by 370
  10. 2020Weight Poisoning Attacks on Pre-trained Modelscited by 341
  11. 2018The Hidden Vulnerability of Distributed Learning in Byzantiumcited by 326
  12. 2021Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networkscited by 323
20 more
  1. 2019Robust Aggregation for Federated Learningcited by 316
  2. 2020Input-Aware Dynamic Backdoor Attackcited by 312
  3. 2020ONION: A Simple and Effective Defense Against Textual Backdoor Attackscited by 262
  4. 2019A backdoor attack against LSTM-based text classification systemscited by 227
  5. 2021Hidden Killer: Invisible Textual Backdoor Attacks with Syntactic Triggercited by 221
  6. 2017Towards Poisoning of Deep Learning Algorithms with Back-gradient Optimizationcited by 196
  7. 2020Bridging Mode Connectivity in Loss Landscapes and Adversarial Robustnesscited by 132
  8. 2021Be Careful about Poisoned Word Embeddings: Exploring the Vulnerability of the Embedding Layers in NLP Modelscited by 105
  9. 2021RAP: Robustness-Aware Perturbations for Defending against Backdoor Attacks on NLP Modelscited by 104
  10. 2021CRFL: Certifiably Robust Federated Learning against Backdoor Attackscited by 91
  11. 2020Can Adversarial Weight Perturbations Inject Neural Backdoors?cited by 56
  12. 2019Attack-Resistant Federated Learning with Residual-based Reweightingcited by 40
  13. 2020One-pixel Signature: Characterizing CNN Models for Backdoor Detectioncited by 34
  14. 2019Distributed Training with Heterogeneous Data: Bridging Median- and Mean-Based Algorithmscited by 29
  15. 2021How to Inject Backdoors with Better Consistency: Logit Anchoring on Clean Datacited by 24
  16. 2020Don't Trigger Me! A Triggerless Backdoor Attack Against Deep Neural Networkscited by 19
  17. 2020Scalable Backdoor Detection in Neural Networkscited by 15
  18. 2020Cassandra: Detecting Trojaned Networks from Adversarial Perturbationscited by 15
  19. 2022Backdoor Attacks in Federated Learning by Rare Embeddings and Gradient Ensemblingcited by 9
  20. 2020Noise-Response Analysis of Deep Neural Networks Quantifies Robustness and Fingerprints Structural Malwarecited by 7

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2023: 1; 2024: 1; 2025: 3; 2026: 3.

8 papers.

  1. 2023Federated Learning for Generalization, Robustness, Fairness: A Survey and Benchmarkcited by 98
  2. 2025NLP Security and Ethics, in the Wildcited by 2
  3. 2026Byzantine Machine Learning: MultiKrum and an optimal notion of robustnesscited by 1
  4. 2024A Novel Defense Against Poisoning Attacks on Federated Learning: LayerCAM Augmented with Autoencodercited by 0
  5. 2025ATR-Bench: A Federated Learning Benchmark for Adaptation, Trust, and Reasoningcited by 0
  6. 2025FedStrategist: A Meta-Learning Framework for Adaptive and Robust Aggregation in Federated Learningcited by 0
  7. 2026Byzantine-Resilient Federated Learning via QUBO-Based Client Selection on Quantum Annealerscited by 0
  8. 2026FedLNS: Leverage LayerNorm Signature Modeling to Mitigate Adversarial Manipulation in Federated LLMscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024SDBA: A Stealthy and Long-Lasting Durable Backdoor Attack in Federated Learning
  2. 2026Bypassing Krum: Selection-Aware Backdoor Attacks in Federated Learning
  3. 2026BackDFL: A Unified Benchmark For Backdoor Attacks and Defenses In Decentralized Federated Learning
  4. 2025Hammer and Anvil: Toward a Theory of Backdoors in Federated Learning
  5. 2022A Knowledge Distillation-Based Backdoor Attack in Federated Learning
  6. 2024Byzantine-Robust Federated Learning: An Overview With Focus on Developing Sybil-based Attacks to Backdoor Augmented Secure Aggregation Protocols
  7. 2025BackFed: An Efficient & Standardized Benchmark Suite for Backdoor Attacks in Federated Learning
  8. 2023FedGrad: Mitigating Backdoor Attacks in Federated Learning Through Local Ultimate Gradients Inspection