- Rating
- 2(−1)
- Confidence
- 4
- Informativeness
- 3
- Presentation
- 3
- Soundness
- 1
Read review
Paper summary
This paper studies how server-side optimizer state, specifically the FedAvgM momentum buffer, affects backdoor persistence in federated learning after malicious clients leave. Using causal state interventions, the authors start from the same attacked checkpoint and compare branches that preserve the attacked momentum, reset it to zero, or swap in momentum from a matched clean run. Training then continues with only benign clients. Even with identical model weights, these branches diverge in attack success rate while clean accuracy stays nearly the same: an average gap of about 8.8 points on ResNet-18/CIFAR-10, and up to 86 points after 20 rounds on Llama-2-7B/SST-2. A simple proposition explains how a decaying momentum difference can leave a lasting parameter displacement. Building on this, the authors propose DynaPhantom, a history-aware attack that blends each malicious update with an exponential moving average of the attacker’s own past submissions and shifts more attack strength toward the end of the attack window. This raises post-attack ASR well above Neurotoxin and other baselines. The paper also shows that restoring a checkpoint with or without its optimizer state leads to different security trajectories, and it suggests resetting optimizer state after malicious updates are detected as a simple defense.
Critical strengths
- The motivation of this paper is insightful. The author also leads a pilot study for observation.
- The writing of the paper is well-constructed. I appreciate the author's writing, which is easy enough but strikes at the key.
- The author evaluates the experiment on both image classification and language modeling.
Critical weaknesses
Strong Weakness:
- Compared methods, such as Neurotoxin, are outdated and not enough. More methods should be considered, such as Batman [1].
- A larger dataset should be considered. The author only experimented on CIFAR-10 with ResNet-10.
- The case study in federated language-model adaptation is also not enough. The author only compared on SST-2 with LLaMA. In the GLUE dataset, there are many NLP classification tasks. So only choosing one task is not enough to validate generalization for DynaPhantom.
- The two components of DynaPhantom (EMA history and late-stage schedule) are never ablated separately, so it’s unclear which one drives the gain.
- DynaPhantom isn’t evaluated against any robust aggregation defense.
Minor Weakness:
- Results from the pilot study can be shown in the introduction section for easy reading.
- The threat model setting can be more detailed. Moreover, key attack settings are missing (trigger design, number of malicious clients, attack rounds, λ, the late-stage schedule, and how “matched budgets” are defined)
[1] Batman: Benign Knowledge Alignment Through Malicious Null Space in Federated Backdoor Attack, CVPR'26.
Questions for authors
- Figure 3 (65.35) and Table 2 (88.43) don’t match, though both report donor ASR as 35.42. Which attack does Figure 3 use?
- The clean-donor branch is said to be matched by “lineage, defense, and training round”. Which defense is used here? It isn’t described anywhere in the paper.