A two-stage, pipeline-co-executed data selection framework with a class-variance-aware importance sampling scheme improves accuracy and cuts training time on real edge hardware.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
A Two-Stage Data Selection Framework for Data-Efficient Model Training on Edge Devices
A two-stage, pipeline-co-executed data selection framework with a class-variance-aware importance sampling scheme improves accuracy and cuts training time on real edge hardware.