Autonomous learning machine for big online data analytics

Abstract

Deep learning became very popular recently, as it is capable of tackling problems that humans are unable to solve through traditional programming. In this paper, we will go over the implementation of a deep learning model known as Autonomous Deep Learning(ADL), which is capable of being fully flexible and has a self-growing network that adapts to the demand of its dataset by growing its hidden layers and nodes. This is especially important in a non-stationary dataset where data sources can come from different origins.Next, we convert ADL to fit in the context of regression, which is called Autonomous Learning Machine (ALM). Afterwards, we will use it to analyze real-world data of the remaining useful life of aircraft gas turbine engines. Lastly, the experimental results will be summarised and compared with results from other machine learning models.