Natural Language Processing
Abstract
In the modern age of information and analytics, natural language processing (NLP) is one of the most important technologies out there. Making sense of complex structures in language and deriving insights and actions from it are crucial from an artificial intelligence perspective. In several domains, the importance of natural language processing is paramount and ever growing, as digital information in the form of language is ubiquitous. Applications of natural language processing include language translation, sentiment analysis, web search applications, customer service automation, text classification, topic detection from text, language modeling, and so forth. Traditional methods of natural language processing relied on the bag of word models, the vector space of word model, and on-hand coded knowledge bases and ontologies. One of the key areas for natural language processing is the syntactic and semantic analysis of language. Syntactic analysis refers to how words are grouped and connected in a sentence. The main tasks in syntactic analysis are tagging parts of speech, detecting syntactic classes (such as verbs, nouns, noun phrases, etc.), and assembling sentences by constructing syntax trees. Semantic analysis refers to complex tasks such as finding synonyms, performing word-verb disambiguation, and so on. As part of this chapter, we will begin with a brief introduction of the traditional natural language processing techniques and then move on to deep learning-based methods involving recurrent neural networks, attention, and transformers.