What is the last technology of NLP
Natural Language Processing
Natural Language Processing (also: Natural Language Programming, in short: NLP dt. Machine processing of natural language) is a technology that aims to enable computers and humans to communicate with one another on an equal footing. NLP combines findings from linguistics with the latest methods in computer science and artificial intelligence.
In order for natural language processing to work, speech recognition must first be worked on. NLP is seen as a promising technology in the HCI area for controlling devices or web applications. For example, the work of chatbots or digital voice assistants was based on this principle.
The development of NLP goes back well into the 1950s, when the scientist Alan Turing published an article entitled "Computing Machinery and Intelligence". In it he presented a method to measure artificial intelligence. The so-called "Turing Test" still exists today.
As early as 1954, researchers had already succeeded in translating sixty sentences in Russian into English with the help of a machine. Euphoric by this start, many computer researchers thought that machine translation was only a matter of time. But it still had to take until the 1980s for the first systems for statistical-based machine translation to be further developed. In the meantime, some approaches have been found to translate information from the “real” world into computer language.
A major evolutionary step was made in the late 1980s. Because then machine learning became popular. Together with the ever increasing computing power of computers, algorithms for NLP could now be used. One of the pioneers in this field was and is still today the linguist Noam Chomsky. The software company IBM also ensured the increasing further development of natural language processing.
Today, NLP-based computer programs can no longer access manually collected data sets, but are also able to independently analyze text corpora such as websites or spoken language.
NLP is based on the basic idea that any form of language, spoken or written, must first be recognized. However, language is a very complex system of characters. It is not just the individual word that is important, but its connection with other words, entire sentences or facts.
What humans naturally learn from birth, computers have to achieve with the help of algorithms. While humans can fall back on their life experience, the computer must be able to fall back on artificially generated experiences. The challenge for the machine processing of natural language is therefore less in producing language than in understanding it.
How it works 
Modern NLP is based on algorithms, which in turn are based on statistical machine learning. The special thing about it is that computers can not only learn based on previously learned dilemmas, but can also identify problems independently and solve new problem areas on the basis of large corpora of documents. Computers do not learn to find a solution for every problem, but you learn general patterns with the help of which they work on individual questions. This makes NLP a preliminary stage for artificial intelligence.
The big advantage of this method is that the more data they get, the better the computers get. A good example of this is Google's translation function. At the beginning, the project was often smiled at. Today the program is able to translate many different texts and even the spoken word with some degree of fluency.
The “Rank Brain” established by Google also uses the method of natural language processing to deliver the right results even to search queries that have never been asked before. The “interpretation” of inputs is supplemented by artificial intelligence.
Computer programs based on NLP must perform the following tasks:
- Simplify text
- Convert text to spoken language
- Convert spoken language to text
- Understand natural language searches
- Recognize advanced questions and follow-up questions
- Check the plausibility of answers
NLP touches on many individual areas. This includes:
- Information retrieval: in the general processing of information
- Information extraction: for semantic questions
- Speech processing: Speech recognition or text-to-speech functions
Speech recognition as a central task area in NLP depends on many different factors. The most important are briefly summarized here.
- automated summary: The programs must be able to automatically reduce large texts to the essentials.
- Word relationships within sentences: Here it is required of NLP that it recognizes which sentence components are related to each other.
example: I sat in the back seat of the car. In this case, the program must recognize that the back seat belongs to the car.
- Discourse analysis: NLP software must be able to recognize the register of a text (raised, colloquial). The program must also recognize what type of text it is (shopping list, invoice, request).
- machine translation: Programs based on NLP must be able to translate human language into another human language and have to master grammar, semantics and other linguistic sub-areas.
- morphological segmentation: This is the breakdown of a word into its individual components.
- NER (Named Entity Recognition) : An NLP program must recognize whether a text contains proper names for places, people or organizations and it must also be able to assign them. For text output, the program must therefore know whether the words in question are capitalized, even in Western languages.
- Conversion into human language: Words stored digitally are translated into human language.
- Understand human language
- Optical character recognition (OCR) : This is an image recognition system that can convert images into text, as some scanners can do today.
- Recognition of feelings
- Recognize spoken language
- Recognizing styles such as irony
- Recognize word meanings: In terms of sound, “booking” can include the action of buying a ticket as well as the majority of the tree “beech”.
Areas of application and outlook
NLP is an important building block in the development of artificial intelligence. Because language plays a central role in the creation of computers that think independently. The approach of natural language processing thus formed the important interface between human beings and computers.
Today these techniques are used for the translation of documents, for the processing of documents, but also for call centers. There are now programs that can create texts independently.
Services such as Skype should soon be able to do so. Translate phone calls live. Even today, users can “talk” to chatbots from selected providers on Skype in order to book tickets or to start simple queries. Google also wants to turn its translator into a live translator. At the same time, the technology is used by numerous digital assistants from large Internet companies, for example at Amazon Echo, Windows Cortana or Siri from Apple.
- ↑ Skype goes Star Trek: Microsoft announces universal translator for 40 languages t3n.de Retrieved on September 26, 2014
- ↑ Google plans real-time translation for Android futurezone.at. Accessed on September 26, 2014
Web links 
- Why do two syllable names sound better
- How to say Find in Japanese
- What is the closest airport to Ooty
- How do I accept my father's death
- What is chemical engineering
- Do we need reasons for our beliefs?
- Counts math as an art
- Which countries have good relations with Sweden?
- Is traceroute a TCP or UDP
- Democracy would lead to political chaos
- What is the astronaut death rate
- How ants feel organs
- What is life's ultimate dilemma
- How to pronounce Kanekalon
- How do I prepare for MIT
- How do you think and live creatively
- Do-off policy guidelines gradient methods are in place
- What fascinates people about K-pop culture
- How can I wait for forgiveness
- How do I send urls to google
- What is the difference between Monetate and MonetiZe
- How can Hermione stand Ron
- How is summer in Paris, France
- What is hydrogen peroxide absorption treatment