RSAM: Byzantine-Robust and Secure Model Aggregation in Federated Learning for Internet of Vehicles Using Private Approximate Median
Bibliographic record
Abstract
In Internet-of-Vehicles (IoVs), Federated Learning (FL) is increasingly used by smart vehicles to process various sensing data. FL is a collaborative learning approach that enables vehicles to train a shared machine learning (ML) model by exchanging their local models instead of their sensitive training data in a distributed manner. Secure aggregation, as a privacy primitive for FL, aims to further protect the local models.c However, existing secure aggregation methods for FL in IoVs mostly suffer from poor security against Byzantine attacks, e.g., malicious vehicles submit fake local models, which are common in IoVs and greatly degrade the accuracy of the final shared model without being detected. In this paper, we propose a new secure and efficient aggregation approach, RSAM, for resisting Byzantine attacks FL in IoVs. RSAM first securely calculates an approximate median of local models of the distributed vehicles via the divide-and-conquer strategy as the aggregation model in each training round, providing the strong Byzantine robustness that is similar to the real median (a proven robust rank-based statistic) does, where median means the coordinate-wise median. Furthermore, RSAM is a single-server secure aggregation protocol that protects the vehicles' local models and training data against inside conspiracy attacks based on zero-sharing. Finally, RSAM is efficient for vehicles in IoVs, since RSAM transforms the sorting operation over the encrypted data to a small number of comparison operations over plain texts and vector-addition operations over ciphertexts, and the main building block relies on fast symmetric-key primitives. The correctness, Byzantine resilience, and privacy protection of RSAM are analyzed, and extensive experiments demonstrate its effectiveness.
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 | 0.002 | 0.006 |
| 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 | 0.002 | 0.003 |
| Open science | 0.002 | 0.004 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 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".