Deep Learning for Technical Document Classification

In large technology companies, the requirements for managing and organizing technical documents created by engineers managers have increased dramatically in recent years, which has led to a higher demand more scalable, accurate, automated document classification. Prior studies only focused on processing text classification, whereas often contain multimodal information. To leverage information classification improve model performance, this article presents novel deep learning architecture, i.e., TechDoc, utilizes three types of information, including natural language texts descriptive images within associations among documents. The architecture synthesizes convolutional neural network, recurrent graph network through an integrated training process. We applied database trained classifying based hierarchical International Patent Classification system. Our results show that TechDoc greater accuracy than unimodal methods other state-of-the-art benchmarks. can potentially be scaled millions real-world documents, is useful data knowledge management companies organizations.