Research map / Federated learning and privacy / Privacy attacks and secure learning
DoBlock: Blocking Malicious Association Propagation for Backdoor-Robust Federated Learning Under Domain Skew
Zhou Tan, De Li, Yirui Huang, Duanshu Fang, Jia-Li Yin, Xiaolei Liu, Songze Li, Shouling Ji
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| keep local models out of aggregation and share only a domain infuser | federated learning under domain skew · Digits, CBA attack | average ASR: 99.39 -> 0.72 (vs FedAvg) | Improved |
| add FiLM and domain classifier to the domain infuser (ablation) | DoBlock · Digits, CBA attack | MA: 87.37 -> 90.19 | Improved |
| add domain classifier Q to infuser with FiLM (ablation) | DoBlock · Digits, CBA attack | ASR: 1.11 -> 0.72 | Improved |
| add domain classifier Q to infuser with FiLM (ablation) | DoBlock · PACS, CBA attack | ASR: 2.42 -> 2.22 | No clear effect |
| inject triggers without changing labels (ablation) | backdoor attack on FedAvg · Digits and PACS | ASR: 99.39 -> 1.39 on Digits | Hurt |
| manipulate labels without triggers (ablation) | backdoor attack on FedAvg · Digits and PACS | ASR: 99.39 -> 10.23 (Digits); 100.00 -> 22.01 (PACS) | Hurt |
| apply robust aggregation defenses such as MKrum, DnC, FLAME, RFA under domain skew | existing defenses · Digits and PACS, CBA | ASR | No clear effect |
| vary attack type (FCBA, SIG, ModRep, Sybil, Neurotoxin, Clean Label) | DoBlock · Digits and PACS | ASR: ASR below 2.45% | No clear effect |
| increase number of clients (ablation) | DoBlock · client scale experiment | MA and RA | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- LockDown — Lockdown: Backdoor Defense for Federated Learning with Isolated Subspace Training (NeurIPS 2023) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Resisting Backdoor Attacks in Federated Learning via Bidirectional Elections and Individual Perspective (AAAI 2024 · similarity 0.87)
- MARS: A Malignity-Aware Backdoor Defense in Federated Learning (NeurIPS 2025 · similarity 0.87)
- SPMC: Self-Purifying Federated Backdoor Defense via Margin Contribution (ICML 2025 · similarity 0.87)
- RFLPA: A Robust Federated Learning Framework against Poisoning Attacks with Secure Aggregation (NeurIPS 2024 · similarity 0.87)
- Parameter Disparities Dissection for Backdoor Defense in Heterogeneous Federated Learning (NeurIPS 2024 · similarity 0.87)
- Detecting Backdoor Attacks in Federated Learning via Direction Alignment Inspection (CVPR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
