Personalized AI apps
Build multi-agent systems without code and automate document search, RAG and content generation
Start free trial Question
Batch Normalization - Why does batch normalization reduce overfitting?
Answer
Batch normalization preserves more information in the weights by allowing the network to use lower weight decay values without significantly diminishing the regularization impact. The network becomes less prone to co-adaptation and overfitting as a result of reduced feature correlation.
