A multi-level cloud-edge-end inference system combining early exit, attention-based pruning, and confidence-based offloading reduces BERT sentiment-analysis latency by up to 17%, but only at measurable accuracy cost.
Learning layer-skippable inference network,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.NI 1years
2024 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Distributed Collaborative Inference System in Next-Generation Networks and Communication
A multi-level cloud-edge-end inference system combining early exit, attention-based pruning, and confidence-based offloading reduces BERT sentiment-analysis latency by up to 17%, but only at measurable accuracy cost.