Speedup of Network Training Process by Eliminating the Overshoots of Outputs - Artificial Intelligence Applications and Innovations (AIAI 2018) Access content directly
Conference Papers Year : 2018

Speedup of Network Training Process by Eliminating the Overshoots of Outputs

Di Zhou
  • Function : Author
  • PersonId : 1033495
Yuxin Zhao
  • Function : Author
  • PersonId : 1033496
Chang Liu
  • Function : Author
  • PersonId : 1033497
Yanlong Liu
  • Function : Author
  • PersonId : 1033498

Abstract

The overshoots between the expected and actual outputs while training network will slow down the training speed and affect the training accuracy. In this paper, an improved training method for eliminating overshoots is proposed on the basis of traditional network training algorithms and a suggestion of eliminating overshoot is given. Gradient descent is regarded as the training criterion in traditional methods which neglects the side effects caused by overshoots. The overshoot definition (OD) is combined with gradient descent. According to the overshoot suggestion, local linearization and weighted mean methods are used to adjust the parameters of network. Based on the new training strategy, a numerical experiment is conducted to verify the proposed algorithm. The results show that the proposed algorithm eliminates overshoots effectively and improves the training performance of the network greatly.
Fichier principal
Vignette du fichier
467708_1_En_39_Chapter.pdf (263.5 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01821056 , version 1 (22-06-2018)

Licence

Attribution

Identifiers

Cite

Di Zhou, Yuxin Zhao, Chang Liu, Yanlong Liu. Speedup of Network Training Process by Eliminating the Overshoots of Outputs. 14th IFIP International Conference on Artificial Intelligence Applications and Innovations (AIAI), May 2018, Rhodes, Greece. pp.462-470, ⟨10.1007/978-3-319-92007-8_39⟩. ⟨hal-01821056⟩
58 View
54 Download

Altmetric

Share

Gmail Facebook X LinkedIn More