Home Publications Research Grants Inventions & Patents Awards Additional Research Activities Faculties & Institutions Research Areas

ForestNet – Automatic Design of Sparse Multilayer Perceptron Network Architectures Using Ensembles of Randomized Trees

Rodriguez Salas D, Ravikumar N, Seuret M, Maier A (2020)

---

Publication Language: English

Publication Type: Conference contribution, Conference Contribution

Publication year: 2020

Publisher: Springer International Publishing

Conference Proceedings Title: Pattern Recognition

Event location: Auckland, New Zealand

ISBN: 978-3-030-41404-7

URI: https://link.springer.com/chapter/10.1007/978-3-030-41404-7_3

DOI: 10.1007/978-3-030-41404-7_3

Abstract

In this paper, we introduce a mechanism for designing the architecture of a Sparse Multi-Layer Perceptron network, for classification, called ForestNet. Networks built using our approach are capable of handling high-dimensional data and learning representations of both visual and non-visual data. The proposed approach first builds an ensemble of randomized trees in order to gather information on the hierarchy of features and their separability among the classes. Subsequently, such information is used to design the architecture of a sparse network, for a specific data set and application. The number of neurons is automatically adapted to the dataset. The proposed approach was evaluated using two non-visual and two visual datasets. For each dataset, 4 ensembles of randomized trees with different sizes were built. In turn, per ensemble, a sparse network architecture was designed using our approach and a fully connected network with same architecture was also constructed. The sparse networks defined using our approach consistently outperformed their respective tree ensembles, achieving statistically significant improvements in classification accuracy. While we do not beat state-of-art results with our network size and the lack of data augmentation techniques, our method exhibits very promising results, as the sparse networks performed similarly to their fully connected counterparts with a reduction of more than 98% of connections in the visual tasks.

Authors with CRIS profile

How to cite

APA:

Rodriguez Salas, D., Ravikumar, N., Seuret, M., & Maier, A. (2020). ForestNet – Automatic Design of Sparse Multilayer Perceptron Network Architectures Using Ensembles of Randomized Trees. In Pattern Recognition. Auckland, New Zealand, NZ: Springer International Publishing.

MLA:

Rodriguez Salas, Dalia, et al. "ForestNet – Automatic Design of Sparse Multilayer Perceptron Network Architectures Using Ensembles of Randomized Trees." Proceedings of the The 5th Asian Conference on Pattern Recognition, Auckland, New Zealand Springer International Publishing, 2020.

BibTeX: Download