Boundary loss plus selective gradient clipping lets polynomial neural networks train stably at high degrees and match ReLU accuracy on seven datasets.
A methodology for training homomorphic encryption friendly neural networks
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 Training Framework for Optimal and Stable Training of Polynomial Neural Networks
Boundary loss plus selective gradient clipping lets polynomial neural networks train stably at high degrees and match ReLU accuracy on seven datasets.