Trustworthy Aggregation for Aerial Federated Learning in Heterogeneous Client Environments
Bibliographic record
Abstract
The integration of unmanned aerial vehicles (UAVs) within Federated Learning (FL) marks a crucial advancement, introducing a dynamic and flexible dimension to decentralized machine learning paradigms. However, the location-dependent performance, characterized by variations in transmission rates and susceptibility to errors, presents challenges for FL's convergence speed and accuracy. This paper proposes a trustworthy aggregation approach implemented on non-terrestrial networks, addressing client heterogeneity in aspects of trustworthiness, available datasets, and transmission rates. The proposed approach initially prioritizes clients with high data rates is implemented to expedite convergence speed, gradually expanding to incorporate a more extensive client base. This accommodation of clients is pivotal for training the model on a larger decentralized dataset that is a crucial consideration in scenarios with non-independent and identically distributed datasets. However, the potential presence of imperfect local models stemming from small datasets, undetected attacks, or less powerful computational devices may result in a gradual degradation of training performance over time. In such instances, the system exclusively focuses on training with a reliable set of clients. The proposed algorithm is subjected to benchmarking against two scenarios: an aggressive approach, which involves accommodating all clients, and a conservative approach, which only includes authenticated clients. Notably, the proposed algorithm excels in both cases, especially in environments with lower levels of trust.
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 | 0.006 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| 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".