Deep Learning with Kernel Regularization for Visual Recognition Poster ID: M51
Kai Yu Wei Xu Yihong Gong

Introduce an informative regularizer for deep models. Transfer prior knowledge from human-designed kernels. Formulate into a stochastic optimization algorithm. The trained deep model is: ­ faster than SVMs in the test time. ­ more accurate than CNNs and SVMs.
·First ·Prev ·Page 1 ·Next ·Last ·Go Back ·Full Screen ·Close ·Quit