PACT: A Passive Accuracy-Based Trust Metric for Malicious Client Detection in Federated Learning
Bibliographic record
Abstract
Detecting malicious clients in federated learning is particularly challenging when adversaries form the majority. Existing defenses often rely on auxiliary validation data, trusted reference clients, or restrictive attacker assumptions that rarely hold in practice. We introduce Passive Accuracy-Based Client Trust (PACT), a fully passive detector that operates without any probe or validation dataset. PACT estimates client trustworthiness by measuring the class-specific accuracy degradation each client update induces in the global model and by combining the mean and standard deviation of these degradations into a single trust score. Clients whose scores fall below a data-driven threshold, determined using Youden’s J statistic, are flagged as malicious. Experiments on MNIST, Fashion-Mnist, and CIFAR-10 demonstrate that PACT surpasses state-of-the-art baselines including Multi-Krum, FoolsGold, LFighter, and FedDMC under adversary ratios up to 80% in targeted label-flipping attacks. Runtime analysis shows that PACT introduces only modest computational overhead. Overall, PACT provides a practical and robust solution for adversary detection in federated learning deployments lacking auxiliary data or trusted participants.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.007 | 0.040 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.003 | 0.007 |
| Open science | 0.003 | 0.004 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".