The abundance of textual information is becoming accessible on the web, paving the way for the prerequisite of techniques and tools to extract meaningful information. One such significant information extraction task is Named Entity Recognition.
Recently, deep learning techniques have been employed in named entity recognition to decrease the cost of human engineering for designing domain-specific features and rules. Deep learning owns the capability to impart continuous real-valued vector representations and semantic composition via nonlinear processing.
Popular applied deep learning frameworks for named entity recognition are Neural Attention, Deep Reinforcement Learning, Deep Adversarial Learning, Deep Active Learning, Deep Transfer Learning, and Deep Multi-task Learning. Information Extraction, Text Clustering, Question-Answering, Information Retrieval, Machine Translation, Knowledgebase or Ontology Population, Automatic Text Summarization, Opinion Mining, And Semantic Search are some of the impressive natural language processing applicative tasks using named entity recognition.
Joint Named Entity Recognition and Entity Linking, an Easy-to-use Toolkit for Deep Learning-based Named Entity Recognition, Scalability of Deep Learning-based Named Entity Recognition, and Deep Learning-based Named Entity Recognition on Informal Text with Auxiliary Resource and Fine-grained Named Entity Recognition and Boundary Detection is the future research scopes of named entity recognition.
Some of the challenges in named entity recognition are Nested entities, Ambiguity in text, Data Annotation, Informal Text, and Unseen Entities and Lack of resources. Several surveys and reviews have been published in named entity recognition, such as deep learning frameworks, recent advancements, current challenges, state-of-the-art practices, natural language processing methods and future research direction, application domains, opportunities, and evaluation measures.