Purpose: This paper presents a novel hash learning-based method for automatic retrieval of English resource database. The problem of fast and accurate retrieval is addressed by analyzing the similarities among words in the database, which are captured by their corresponding hashes. These word hashes are also called “clouds”. They provide a compact representation of the vocabulary of source database. The main advantage of this approach is that it can exploit the statistical information in cloud to improve retrieval performance. The proposed method iteratively computes "clouds" by hashing a search query and then computing their similarities with clouds from similar databases. The similarity scores between “clouds” are computed at each iteration, which will be utilized to update the weights associated with these clouds from each database. The statistics analysis results indicated: Education by the AR technology has a positive and meaningful impact on the LM in components of learning facilities, family support and participation, peer influence, student personality, and student CT.