RMDL: Random Multimodel Deep Learning for Classification
This addresses the need for robust classification methods for complex datasets across domains like text and images, though it appears incremental as an ensemble extension of existing deep learning techniques.
The paper tackles the problem of finding optimal deep learning structures for classification by introducing Random Multimodel Deep Learning (RMDL), an ensemble approach that improves robustness and accuracy across diverse data types, achieving consistently better performance than standard methods on datasets like MNIST and IMDB.
The continually increasing number of complex datasets each year necessitates ever improving machine learning methods for robust and accurate categorization of these data. This paper introduces Random Multimodel Deep Learning (RMDL): a new ensemble, deep learning approach for classification. Deep learning models have achieved state-of-the-art results across many domains. RMDL solves the problem of finding the best deep learning structure and architecture while simultaneously improving robustness and accuracy through ensembles of deep learning architectures. RDML can accept as input a variety data to include text, video, images, and symbolic. This paper describes RMDL and shows test results for image and text data including MNIST, CIFAR-10, WOS, Reuters, IMDB, and 20newsgroup. These test results show that RDML produces consistently better performance than standard methods over a broad range of data types and classification problems.