Exploiting the potential of deep reinforcement learning for classification tasks in high-dimensional and unstructured data
This addresses the challenge of feature selection in large unstructured datasets for machine learning practitioners, though it appears incremental as it builds on existing RL methods.
The paper tackles the problem of feature selection for classification on high-dimensional unstructured data by proposing a framework that combines a Deep Convolutional Autoencoder with reinforcement learning algorithms like Double DQN and Retrace, resulting in efficient policies that use fewer features to achieve high classification precision.
This paper presents a framework for efficiently learning feature selection policies which use less features to reach a high classification precision on large unstructured data. It uses a Deep Convolutional Autoencoder (DCAE) for learning compact feature spaces, in combination with recently-proposed Reinforcement Learning (RL) algorithms as Double DQN and Retrace.