Proposes FL+DP for offloading time-series transformer training in 6G vehicles, claiming 75% accuracy at ε=0.8, but lacks the described algorithms and has incorrect composition accounting.
Machine learning applications in cr-iov,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Privacy-Preserving Offloading for Large Language Models in 6G Vehicular Networks
Proposes FL+DP for offloading time-series transformer training in 6G vehicles, claiming 75% accuracy at ε=0.8, but lacks the described algorithms and has incorrect composition accounting.