An adaptive DNN partitioning framework for heterogeneous edge-cloud systems reduces energy consumption by 27-36% and end-to-end latency by 6-23% versus static baselines on real hardware with VGG16, AlexNet, and MobileNetV2.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.DC 1years
2026 1verdicts
UNVERDICTED 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Adaptive DNN Partitioning and Offloading in Heterogeneous Edge-Cloud Continuum
An adaptive DNN partitioning framework for heterogeneous edge-cloud systems reduces energy consumption by 27-36% and end-to-end latency by 6-23% versus static baselines on real hardware with VGG16, AlexNet, and MobileNetV2.