ATM-Net reports up to 96.93% accuracy and 87.5% power reduction by combining early exits with quantized precision, though the accuracy figure is Q8 not Q4 and the end-to-end system is never measured.
Deep residual learning for image recognition,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
ATM-Net: Adaptive Termination and Multi-Precision Neural Networks for Energy-Harvested Edge Intelligence
ATM-Net reports up to 96.93% accuracy and 87.5% power reduction by combining early exits with quantized precision, though the accuracy figure is Q8 not Q4 and the end-to-end system is never measured.