Mitigating Poisoning Attack Coalitions in FL: A Non-Linear Personalized Approach
Bibliographic record
Abstract
Federated learning (FL) was introduced to mitigate security risks in traditional machine learning (ML) and deep learning (DL), particularly concerning data privacy (DP), through its decentralized architecture. However, FL remains susceptible to cyber-security threats. Such risks arise when several legitimate client nodes are transformed into a zombie network. These compromised nodes can launch large-scale attacks that affect the confidentiality, integrity, and availability (CIA) of the FL network, ultimately leading to a decline in prediction metrics for legitimate clients. While previous studies have focused on specific, small-scale attacks, none have addressed the impact of large-scale attacks, specifically, an "attack coalition," where over 50% of the network is compromised, and multiple types of attacks occur simultaneously. Many existing studies rely on complex methods such as blockchain or malicious node detection, and only a few have considered the simpler alternative of "personalization," though with limitations. In this paper, we introduce the concept of attack coalition (AC), where malicious client nodes coordinate to undermine the FL system’s performance. We analyze the impact of this AC and propose a personalized FL approach to mitigate its effects. To validate our approach, we use a publicly available dataset measuring nurse stress levels, implemented through a deep neural network (DNN). Our results demonstrate that the AC severely reduces prediction metrics, with accuracy and F1-scores dropping to 18.34% and 18.68%, respectively. By applying our proposed method, these metrics improve significantly, reaching 99.36% and 98.46%, respectively.
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 | 0.003 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.004 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
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".