Soft sparsity via MSB proxy skips negligible multiplications in CNNs, reducing ReLU MACs 88% and Tanh MACs 75% on LeNet-5 with zero accuracy loss.
Akhlaghi et al., ”SnaPEA: Predictive early activation for reducing computation in deep convolutional neural networks,” inProc
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Hardware Efficient Approximate Convolution with Tunable Error Tolerance for CNNs
Soft sparsity via MSB proxy skips negligible multiplications in CNNs, reducing ReLU MACs 88% and Tanh MACs 75% on LeNet-5 with zero accuracy loss.