Lune

EMNLP Findings2022

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

Zhiyuan Zhang, Qi Su, Xu Sun

Abstract

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.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

3 top-tier papers have cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Latent Backdoor Attacks on Deep Neural Networks cites Trojaning Attack on Neural Networks
  • Input-Aware Dynamic Backdoor Attack cites Latent Backdoor Attacks on Deep Neural Networks
  • Input-Aware Dynamic Backdoor Attack cites Trojaning Attack on Neural Networks
  • Fed-FA: Theoretically Modeling Client Data Divergence for Federated Language Backdoor Defense cites Latent Backdoor Attacks on Deep Neural Networks
  • Fed-FA: Theoretically Modeling Client Data Divergence for Federated Language Backdoor Defense cites Trojaning Attack on Neural Networks
  • Fed-FA: Theoretically Modeling Client Data Divergence for Federated Language Backdoor Defense cites Input-Aware Dynamic Backdoor Attack
  • Fed-FA: Theoretically Modeling Client Data Divergence for Federated Language Backdoor Defense cites Hidden Trigger Backdoor Attacks
  • Fed-FA: Theoretically Modeling Client Data Divergence for Federated Language Backdoor Defense cites Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networks
  • Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networks cites Latent Backdoor Attacks on Deep Neural Networks
  • Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networks cites Trojaning Attack on Neural Networks
  • Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networks cites Input-Aware Dynamic Backdoor Attack
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers3

  1. Parameter Disparities Dissection for Backdoor Defense in Heterogeneous Federated LearningWenke Huang, Mang Ye, Zekun Shi, Guancheng Wan et al.NeurIPS 2024 · 12 citations
  2. Fed-FA: Theoretically Modeling Client Data Divergence for Federated Language Backdoor DefenseZhiyuan Zhang, Deli Chen, Hao Zhou, Fandong Meng et al.NeurIPS 2023 · 10 citations
  3. A Practical and Secure Byzantine Robust AggregatorDe Zhang Lee, Aashish Kolluri, Prateek Saxena, Ee-Chien ChangCCS 2025

Builds on13

  1. Trojaning Attack on Neural NetworksYingqi Liu, Shiqing Ma, Yousra Aafer, Wen-Chuan Lee et al.NDSS 2018 · 1,377 citations
  2. Hidden Trigger Backdoor AttacksAniruddha Saha, Akshayvarun Subramanya, Hamed PirsiavashAAAI 2020 · 743 citations
  3. Input-Aware Dynamic Backdoor AttackTuan Anh Nguyen, Anh Tuan TranNeurIPS 2020 · 601 citations
  4. Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural NetworksYige Li, Xixiang Lyu, Nodens Koren, Lingjuan Lyu et al.ICLR 2021 · 548 citations
  5. Latent Backdoor Attacks on Deep Neural NetworksYuanshun Yao, Huiying Li, Haitao Zheng, Ben Y. ZhaoCCS 2019 · 465 citations

Related papers

  1. BadPre: Task-agnostic Backdoor Attacks to Pre-trained NLP Foundation ModelsKangjie Chen, Yuxian Meng, Xiaofei Sun, Shangwei Guo et al.ICLR 2022 · 133 citations
  2. Fine-mixing: Mitigating Backdoors in Fine-tuned Language ModelsZhiyuan Zhang, Lingjuan Lyu, Xingjun Ma, Chenguang Wang et al.EMNLP Findings 2022 · 16 citations
  3. Enhancing Byzantine-Resistant Aggregations with Client EmbeddingZhiyuan Zhang, Hao Zhou, Fandong Meng, Jie Zhou et al.EMNLP Findings 2024 · 2 citations
  4. Backdoor Attacks in Federated Learning by Rare Embeddings and Gradient EnsemblingKiYoon Yoo, Nojun KwakEMNLP 2022 · 24 citations
  5. RAP: Robustness-Aware Perturbations for Defending against Backdoor Attacks on NLP ModelsWenkai Yang, Yankai Lin, Peng Li, Jie Zhou et al.EMNLP 2021 · 57 citations
Dusk over the sea between two cliffs drawn in fine vertical lines