Abstract
This paper introduces a novel hybrid deep neural kernel framework. The proposed deep learning model follows a combination of neural networks based architecture and a kernel based model. In partic- ular, here an explicit feature map, based on random Fourier features, is used to make the transition between the two architectures more straight- forward as well as making the model scalable to large datasets by solving the optimization problem in the primal. The introduced framework can be considered as the first building block for the development of even deeper models and more advanced architectures. Experimental results show a significant improvement over shallow models on several medium to large scale real-life datasets.
Original language | English |
---|---|
Title of host publication | ESANN 2017 - Proceedings, 25th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning |
Publisher | i6doc.com publication |
Pages | 17-22 |
Number of pages | 6 |
ISBN (Electronic) | 9782875870391 |
Publication status | Published - 2017 |
Event | 25th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2017 - Bruges, Belgium Duration: 26 Apr 2017 → 28 Apr 2017 |
Publication series
Name | ESANN 2017 - Proceedings, 25th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning |
---|
Conference
Conference | 25th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2017 |
---|---|
Country/Territory | Belgium |
City | Bruges |
Period | 26/04/17 → 28/04/17 |
Bibliographical note
Funding Information:The authors acknowledge support of ERC AdG A-DATADRIVE-B (290923), KUL: GOA/10/09 MaNet, CoE PFV/10/002 (OPTEC), BIL12/11T; FWO: G.0377.12, G.088114N, G0A4917N; IUAPP7/19 DYSCO.
Publisher Copyright:
© ESANN 2017 - Proceedings, 25th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning. All rights reserved.
Funding
The authors acknowledge support of ERC AdG A-DATADRIVE-B (290923), KUL: GOA/10/09 MaNet, CoE PFV/10/002 (OPTEC), BIL12/11T; FWO: G.0377.12, G.088114N, G0A4917N; IUAPP7/19 DYSCO.