Abstract:
A desired capability of deep learning is to understand the high-level, class-specific features via hierarchical features learning. However the training of deep architectu...Show MoreNotes: As originally published there is an error in this document. The authors omitted the following text:"As the training proceeds, the average inner product and the classification accuracy at each epoch are illustrated in Figure 3 and Figure 4 respectively, with initialisation range of W optimised separately for each scenario." The article PDF remains as originally published.
Metadata
Abstract:
A desired capability of deep learning is to understand the high-level, class-specific features via hierarchical features learning. However the training of deep architectures is costly comparing to simple shallow models. Bringing the high-level feature understanding into a simple shallow architecture remains an open question.
Notes: As originally published there is an error in this document. The authors omitted the following text:"As the training proceeds, the average inner product and the classification accuracy at each epoch are illustrated in Figure 3 and Figure 4 respectively, with initialisation range of W optimised separately for each scenario." The article PDF remains as originally published.
Date of Conference: 12-17 July 2015
Date Added to IEEE Xplore: 01 October 2015
ISBN Information:
Contents 