Skip to main content
placeholder image

Embedding of FRPN in CNN architecture

Conference Paper


Abstract


  • This paper extends the fully recursive perceptron network (FRPN) model for vectorial inputs to include deep convolutional neural networks (CNNs) which can accept multi-dimensional inputs. A FRPN consists of a recursive layer, which, given a fixed input, iteratively computes an equilibrium state. The unfolding realized with this kind of iterative mechanism allows to simulate a deep neural network with any number of layers. The extension of the FRPN to CNN results in an architecture, which we call convolutional-FRPN (C-FRPN), where the convolutional layers are recursive. The method is evaluated on several image classification benchmarks. It is shown that the C-FRPN consistently outperforms standard CNNs having the same number of parameters. The gap in performance is particularly large for small networks, showing that the C-FRPN is a very powerful architecture, since it allows to obtain equivalent performance with fewer parameters when compared with deep CNNs.

Publication Date


  • 2020

Citation


  • Rossi, A., Hagenbuchner, M., Scarselli, F., & Tsoi, A. C. (2020). Embedding of FRPN in CNN architecture. In ESANN 2020 - Proceedings, 28th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (pp. 133-138).

Scopus Eid


  • 2-s2.0-85098960538

Web Of Science Accession Number


Start Page


  • 133

End Page


  • 138

Abstract


  • This paper extends the fully recursive perceptron network (FRPN) model for vectorial inputs to include deep convolutional neural networks (CNNs) which can accept multi-dimensional inputs. A FRPN consists of a recursive layer, which, given a fixed input, iteratively computes an equilibrium state. The unfolding realized with this kind of iterative mechanism allows to simulate a deep neural network with any number of layers. The extension of the FRPN to CNN results in an architecture, which we call convolutional-FRPN (C-FRPN), where the convolutional layers are recursive. The method is evaluated on several image classification benchmarks. It is shown that the C-FRPN consistently outperforms standard CNNs having the same number of parameters. The gap in performance is particularly large for small networks, showing that the C-FRPN is a very powerful architecture, since it allows to obtain equivalent performance with fewer parameters when compared with deep CNNs.

Publication Date


  • 2020

Citation


  • Rossi, A., Hagenbuchner, M., Scarselli, F., & Tsoi, A. C. (2020). Embedding of FRPN in CNN architecture. In ESANN 2020 - Proceedings, 28th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (pp. 133-138).

Scopus Eid


  • 2-s2.0-85098960538

Web Of Science Accession Number


Start Page


  • 133

End Page


  • 138