An analytic formulation of convolutional neural network learning for pattern recognition

Open Webpage

Huiping Zhuang, Zhiping Lin, Yimin Yang, Kar-Ann Toh

Published: 2025, Last Modified: 01 Apr 2026Inf. Sci. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•A fresh analytic formulation (ACnnL) for training CNNs unlike the iterative strategies.•The CNN structure can be treated as a generalized MLP with significantly more constraints.•The ACnnL completes the training with a significantly faster speed than BP.•For small-sample datasets, the ACnnL excels in prediction comparing with that of BP.