Broad learning vs. deep learning

Abstract

Machine Learning has been gaining traction in recent years due to the many successful implementation of it. Major companies are applying Machine Learning to automate processes which has led to an increase in efficiency at work. Deep Learning models in particular have been used in many of these processes. However, it is extremely inefficient for the data scientist to remodel the model as the entire model would have to be retrained before the data scientist can observe the new model. Thus, an alternative to the Deep Learning model would be the Broad Learning model whereby the data scientist would be able to alter the model and observe the remodelled model while the simulation is ongoing. Thus, this project is to draw comparisons between a Deep Learning model, specifically, the Multi-Layered Convolutional Neural Network and the Broad Learning model to determine if the Broad Learning model would be a suitable alternative to the Deep Learning model. The two models would be tested against the MNIST handwritten digits dataset.