©2020-2024 ioDraw All rights reserved
natural language processing (NLP) As AI The Pearl in the world crown , It mainly covers two processes , Generative and natural language understanding , Use a formula to express , Can be expressed as ：NLP=NLU+NLG. Take machine translation as an example , For the translation between two languages , The machine translation model plays an intermediate role , First of all, the model processes one of the languages ( Let's not talk about those abstruse encoding and decoding processes ), This includes work done before converting to another language , It is called natural language understanding ; Through language understanding process , Further generate another language , This process is called natural language generation . In a nutshell ,NLU It refers to the conversion of text or voice to meaning This is the intermediate result .NLG It means by meaning The intermediate result is text or voice . Natural language processing as the core of cognitive intelligence , If language intelligence can make a breakthrough , So the whole research of artificial intelligence will take a big step . Natural language processing is an important technology to embody language intelligence , It can be analyzed , Understanding or generating natural language , Realize the natural communication between human and machine , At the same time, it is also conducive to the communication between people . From the scope of the study to analyze , Natural language processing includes three parts , The first part is NLP The basic research of English , Include participles , Part of speech tagging , Named entity recognition , Syntactic analysis , Semantic analysis ; The second part is on the basis of the core areas of research , Including vocabulary , phrase , sentence , The expression of discourse . There are major areas of machine translation , Information extraction , Chat and conversation , reading comprehension , Language generation , knowledge engineering , Recommender system , Emotional analysis, etc ; The third part is the analysis NLP Application landing , It's about how to NLP Technology goes deep into each application system , For example, intelligent translation machine , Intelligent customer service , Search Engines , Voice assistant , Knowledge Q & A and other large application systems .
Natural language processing has also experienced from the traditional knowledge-based reasoning method to the data-driven shallow machine learning method , Then to the current big data-driven deep learning . Both academic research and engineering landing have made rapid development . With the continuous development of research , There are also many more challenging applications , For example, simultaneous interpretation technology , Unsupervised multilingual translation , Emotional dialogue （ How to improve the rigid dialogue , Make the dialogue more emotional ）, Poetry generation, etc .NLP broad and profound , Although the data-driven approach does work , however , How to combine the feature information of language to increase the interpretability of the model , This is of great significance to further improve the learning algorithm , And it will be NLP The next hot topic .
Technology