What is IOB? Explain ELMo as a deep learning algorithm used in Named Entity Recognition.
IOB or Inside-Outside-Beginning is a tagging format used to tag tokens or chunks. This technique basically informs about the inside, outside and beginning of a chunk and it also allows multiple noun phrases. Here, it indicates that the word is inside the current noun phrase, O indicates the end of a sentence and B indicates the beginning of a noun phrase.
ELMo is a word representation model that is based on word use characteristics and their variation across linguistic contexts. They use bidirectional Language Model (biLM) which is trained before on a large text corpus. They are represented by contextual, deep network or character based.
The above flowchart shows how ELMo is used in a sentence to extract an entity.