RMDL: Random Multimodel Deep Learning for Classification

  • 2018-05-03 19:36:43
  • Kamran Kowsari, Mojtaba Heidarysafa, Donald E. Brown, Kiana Jafari Meimandi, Laura E. Barnes
  • 42

Abstract

The continually increasing number of complex datasets each year necessitatesever improving machine learning methods for robust and accurate categorizationof these data. This paper introduces Random Multimodel Deep Learning (RMDL): anew ensemble, deep learning approach for classification. Deep learning modelshave achieved state-of-the-art results across many domains. RMDL solves theproblem of finding the best deep learning structure and architecture whilesimultaneously improving robustness and accuracy through ensembles of deeplearning architectures. RDML can accept as input a variety data to includetext, video, images, and symbolic. This paper describes RMDL and shows testresults for image and text data including MNIST, CIFAR-10, WOS, Reuters, IMDB,and 20newsgroup. These test results show that RDML produces consistently betterperformance than standard methods over a broad range of data types andclassification problems.

 

Introduction (beta)

None

 

Conclusion (beta)

None