A Survey on Deep Learning for Named Entity Recognition

Named entity recognition (NER) is the task to identify mentions of rigid designators from text belonging predefined semantic types such as person, location, organization etc. NER always serves foundation for many natural language applications question answering, summarization, and machine translation. Early systems got a huge success in achieving good performance with cost human engineering designing domain-specific features rules. In recent years, deep learning, empowered by continuous real-valued vector representations composition through nonlinear processing, has been employed systems, yielding stat-of-the-art performance. this paper, we provide comprehensive review on existing learning techniques NER. We first introduce resources, including tagged corpora off-the-shelf tools. Then, systematically categorize works based taxonomy along three axes: distributed input, context encoder, tag decoder. Next, survey most representative methods applied new problem settings applications. Finally, present readers challenges faced outline future directions area.