SAL is a hierarchical framework that trains deep neural networks by starting with the simplest network and using its hidden and output layers to guide more complex networks, resulting in more stable training and better generalization.
Title resolution pending
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
-
Self-Abstraction Learning for Effective and Stable Training of Deep Neural Networks
SAL is a hierarchical framework that trains deep neural networks by starting with the simplest network and using its hidden and output layers to guide more complex networks, resulting in more stable training and better generalization.