A Recurrent Neural Model for Temporal Information Extraction

Abstract

Temporal information extraction is an emerging area of information retrieval. Understanding temporal nature of a document is very important in application like answering time sensitive queries, doing temporal analysis of a document, document clustering etc. Lot of research is done in temporal reasoning using rule based or machine learning based approaches. In this paper, deep learning is used to extract temporal expressions from the text documents. Bi-directional Long Short term Memory Recurrent Neural Network (Bi-LSTM RNN) is used to extract temporal expression from the text. Gold standard datasets are used for training and evaluation. Performance of the proposed system is compared with existing system.