Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Sergey Ioffe and Christian Szegedy submitted Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift on February 11, 2015. The paper proposed normalizing layer inputs as part of the model architecture using statistics computed for each training mini-batch, and reported that the method enabled higher learning rates and reduced sensitivity to initialization.
Historical classification
This verified record belongs to LINEAiGE's artificial-intelligence history layer beginning in 1956.
Why it matters
A documented deep-learning optimization and normalization milestone introducing Batch Normalization as an architectural training method.
People, organizations & technologies
PEOPLE
Sergey Ioffe, Christian Szegedy
TECHNOLOGIES
Batch Normalization, Deep neural networks, Mini-batch normalization
Evidence & verification
STATE
verified
CONFIDENCE
high
LAST VERIFIED
2026-09-11T15:06:32+08:00
The original arXiv record identifies the authors, February 11 2015 submission date, and the method of normalizing layer inputs for each training mini-batch. The official PMLR ICML 2015 proceedings independently corroborate the authors, title, method description and 2015 conference publication. The day-precision date is attached specifically to the arXiv submission milestone and is not asserted as a separate method-birth date. No geography, institutional affiliation, ancestry, influence, causality or relationships are inferred.
Sources
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
arXiv
PUBLISHED
2015-02-11
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Proceedings of Machine Learning Research
Trace the chronology
Record identity
LINEAiGE IDbatch-normalization-2015