Quick and robust feature selection: the strength of energy-efficient sparse training for autoencoders (Extended Abstract)

Feature selection, which identifies the most relevant and informative attributes of a dataset has been introduced to address the challenges raised by the emerge of high-dimensional data. Most existing feature selection methods are computationally inefficient; inefficient algorithms lead to high energy consumption, which is not desirable for devices with limited computational and energy resources. In [1], a novel and flexible method for unsupervised feature selection is proposed. This method, named "QuickSelection", introduces the strength of the neuron in sparse neural networks as a criterion to measure the feature importance. When tested on several benchmark datasets, the proposed method is able to achieve the best trade-off of classification and clustering accuracy, running time, and memory usage, among widely used approaches for feature selection.