Learning Arithmetic Operations With A Multistep Deep Learning - Archive ouverte HAL Access content directly
Conference Papers Year : 2020

Learning Arithmetic Operations With A Multistep Deep Learning

Bastien Nollet
  • Function : Author
  • PersonId : 1076500

Abstract

Deep neural networks are difficult to train when applied to tasks that can be expressed as algorithmic procedures. In this article, we propose to study how the explicit guidance of a network through all steps of the algorithm, using external memory and active choice of inputs, can improve its learning capability. The idea is to take inspiration from a child's learning and running through a procedure via interaction with an external support such as a paper. We show that this mechanism applied to a simple multilayer perceptron can significantly improve its performance when learning either a multi-digit addition or multiplication, which are simple but yet challenging operations to learn.
Fichier principal
Vignette du fichier
PID6463747.pdf (466.25 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-02929738 , version 1 (22-10-2020)

Identifiers

Cite

Bastien Nollet, Mathieu Lefort, Frédéric Armetta. Learning Arithmetic Operations With A Multistep Deep Learning. The International Joint Conference on Neural Networks (IJCNN), Jul 2020, Glasgow, United Kingdom. ⟨10.1109/IJCNN48605.2020.9206963⟩. ⟨hal-02929738⟩
134 View
1120 Download

Altmetric

Share

Gmail Facebook X LinkedIn More