Year of Graduation
Author Identification in Literature Texts Through Deep Learning
Big Data Systems
The main problem of previous models is hard to scale out and to update when new big data comes. This paper suggests a novel architecture to parallelize previous monolithic deep learning models in author identification task. On the top of the parallelized deep learning models, we build a special layer which reducing searching spaces called Skim-Trigger. It outperforms previous models in terms of training and execution speed.