Recent advances in technology have led to easy data acquisition mechanisms in many fields, leading to massive datasets. It is often of interest to understand the inherent structure and learn the best representation of the given dataset. Graphs are a powerful way to model interrelationships between data features-well constructed meaningful graphs help in representing and processing the data effectively. The graph topology needs to be inferred from the observed data. In this survey, we briefly explore signal processing based graph learning approaches that have been proposed in the literature and propose new research directions. © 2019 IEEE.