Now, artificial intelligence has become a public familiar vocabulary, and natural language processing, but few people understand. A sub-field NLP (Natural Language Processing, NLP) belonging to artificial intelligence, refers to form natural language, sound and meaning information processed by a computer, the input of characters, words, sentences, text output, identification, analysis, understanding, generation and processing operations like. There are many important impact on humans and computers interact.
After thousands of years of human language development has become a subtle form of communication, it carries a wealth of information, which often go beyond the language itself. Natural language processing will become an important technology to communicate with the digital data to fill the human divide. Here are just look at a few common applications of natural language processing:
1, machine translation
With the rapid development of communication technology and Internet technology, as well as a sharp increase in international contact information even more tightly, so that everyone in the world can get across the language barrier challenges of the information is beyond the capabilities of human translation.
Machine translation because of its high efficiency, low cost to meet the needs of countries around the world fast translation of multilingual information. A branch of machine translation is a natural language information processing, it is possible to automatically generate a natural language natural language to another without the need of human help and computer systems. Currently, the Google translation, translation Baidu, Sogou translation industry giants such as artificial intelligence translation platform launched gradually by virtue of its high efficiency and accuracy of the translation process of translation industry occupies a dominant position.
2, combat spam
Currently, the spam filter has become the first line of defense against spam problem. However, many people encounter these problems using e-mail: unwanted e-mail is still being received, or important emails are filtered out. In fact, to determine whether a message is spam, the method first used the "keyword filtering", if there is a common e-mail spam keywords, it is judged as spam. However, this method is very satisfactory, one normal mail may also have these keywords, very easy to misjudge, the second is the keyword deformation, it is easy to circumvent keyword filters.
Natural language processing by analyzing the text content of your message, can be relatively accurately determine whether a message is spam. Currently, Bayesian (Bayesian) spam filtering technology is one of concern, it is by learning a lot of spam and non-spam, mail collection feature words of garbage generated and non-spam thesaurus thesaurus, then according to probability and statistics to calculate the frequency of these lexicon of the spam, in order to be judged.
3, information extraction
Many important decisions in financial markets are increasingly out of human supervision and control. Algorithmic trading is becoming increasingly popular, which is a form of financial investment is fully controlled by the technology. However, many of these financial decisions are affected by the news. Therefore, a major task of natural language processing is to obtain these express proclamation, and extract the relevant information in a format that can be incorporated into algorithmic trading decisions. For example, the merger of messages between trading companies could have a significant impact on the decision to merge the details (including participants, the purchase price) into the trading algorithm, or it will affect the profits of millions of dollars.
4, sentiment analysis
In the digital age, information overload is a real phenomenon, our ability to acquire knowledge and information has far exceeded our ability to understand it. Moreover, this trend no signs of slowing down, so the ability to sum up the meaning of the documents and information is becoming increasingly important. Sentiment analysis as a common method of application of natural language processing, allowing us to identify and absorb relevant information from large amounts of data, but also understand the deeper meaning. For example, companies analyze consumer feedback on the product, or to detect the difference in assessment of information online reviews.
5, QA
With the rapid development of the Internet, the network is increasing the amount of information that people need to get more accurate information. Traditional search engine technology can not meet the increasingly high demand, and automatic question answering technology as an effective means to solve this problem. QA refers to the task automatically by a computer user to answer questions raised in order to meet the knowledge needs of users, in answering user questions, we must first correctly understand the issues raised by the user, to extract critical information which, in the existing corpus of knowledge or carried out the library search, match, will get the answer back to the user.
6, personalized recommendation
Natural language processing can be recorded based on big data and historical behavior, learning the user's interests, predict user ratings for a given article or preferences, to achieve precise understanding of user intent, while the language matching calculation for precise matching. For example, in the field of information services through users to read the content, duration, comments, and other preferences, as well as social networks or even mobile device models used, etc., a comprehensive analysis of user interest information source and core vocabulary, professional refinement analysis so as to perform newsfeeds to achieve personal customized news services, and ultimately enhance the user stickiness.
Written in the last:
Target natural language processing that make up human communication (natural language) with the understanding gap between the computer (machine language), and ultimately the computer to understand natural language like a human intelligence. The future, the development of artificial intelligence, natural language processing will be gradually face more complex situations and solve more problems, will also bring a more intelligent era for us.
from：https://baijiahao.baidu.com/s?id=1641747657949904097&wfr=spider&for=pc
