Convolutional Neural Networks for Literature Retrieval
Qin Yaxue · 2020
Recently, the amount of literature has increased rapidly. The traditional literature retrieval technology can no longer meet the requirements of massive document retrieval. With the rapid development of deep learning (especially convolutional neural networks), many results have been achieved in image, video and natural language processing. Here we proposed a literature retrieval framework with convolutional neural network. Specifically, we first use word vector representation method to transform text. Then we input text into the convolutional neural network with hash layer to extract textual hash features and complete the literature retrieval. Experimental results show that the literature retrieval framework with convolutional neural network improves the accuracy of retrieval, and proves that the application of convolutional neural network in document retrieval can simplify the retrieval process and improve the document retrieval system.