Talking about the Application Field of Natural Language Processing Technology

Text classification Text ClassificaTIon

Text categorization refers to giving a text a prediction of the predetermined category to which it belongs.

"The purpose of text categorization is to categorize the subject or subject of the document."

- p575, FoundaTIons of StaTIsTIcal Natural Language Processing (http://amzn.to/2ePBz9t), 1999

One popular text categorization application is sentiment analysis (https://en.wikipedia.org/wiki/Sentiment_analysis), where common category labels represent emotional tones of the source text, such as "positive" or "negative".

Other three types of text classification applications:

Spam filtering, which sorts emails by email based on text.

Language recognition, which classifies the language of the source text.

The classification of themes, the genre of the classified fictional story.

In addition, you can assign multiple category labels (so-called multi-label classifications) to the text as needed. For example, assign multiple topic tags to a tweet.

Language modeling is really a sub-task of a very interesting natural language problem, especially on the basis of other tasks.

"The problem is to predict the next word for a given word. This task is the basis for speech or optical character recognition, and is also used for spell correction, handwriting recognition, and statistical machine translation."

- p575, Foundations of Statistical Natural Language Processing (http://amzn.to/2ePBz9t), 1999.

In addition to interest in academic research, language models are a key component of many natural language processing architectures that apply deep learning.

The language model can learn the probability relationship between words and words, and then generate a new sequence of words that is statistically consistent with the source text.

The language model can be used for text or speech generation, as follows:

Generate a new article title.

Generate new sentences, paragraphs or documents.

Generate subsequent sentence suggestions.

Speech recognition is the solution to how to understand what humans are saying.

"The task of speech recognition is to convert the acoustic signals of natural language, including spoken language, into a sequence of words that conform to the speaker's expectations."

- p458, Deep Learning (http://amzn.to/2uE7WvS), 2016.

Given the audio data generated from the text, the model must be able to generate human readable text. Given the autonomy of the process, this task can also be called Automatic Speech Recognition (ASR).

The language model is used to create output text based on audio data. Applications include:

Generate a speech text.

Create captions for movies or TV shows.

Send a command to the radio while driving.

Description generation is a problem of how to describe image content, in accordance with digital image generation such as photographs and text descriptions related to image content.

Explain that the generated language model is used to generate a title based on the image. Some specific applications include:

Describe the content of the scene

Create a photo title

Description video

Machine translation is the conversion of source text in one language to another.

"Machine translation, the automatic translation of text or speech from one language to another, is one of the most important applications of NLP."

- p463, Foundations of Statistical Natural Language Processing (http://amzn.to/2ePBz9t), 1999.

Given the inclusion of deep neural networks, this task is now also known as neural machine translation.

"In machine translation tasks, input is composed of a sequence of symbols in a language, and computer programs must convert input into sequences of symbols in other languages. Machine translation is often applied to natural languages, such as from English to French. Recently, depth Learning begins to have a major impact on the task."

- p98, Deep Learning (http://amzn.to/2uE7WvS), 2016

The language model of machine translation is used to output the target text of the second language based on the source text.

Document summarization refers to the task of creating a corresponding short description based on text. Its language model is used to output a summary based on a complete document.

Related applications are as follows:

· Create a document title.

· Generate a document summary.

Answering a question means giving a topic (such as a text document) to answer specific questions about the topic.

“Q&A system, which attempts to answer user queries presented in question form by returning the corresponding phrase (such as location, person or date). For example, why did the problem kill President Kennedy? Maybe the noun phrase Oswald is the answer”

- p377, Foundations of Statistical Natural Language Processing (http://amzn.to/2ePBz9t), 1999.

Common applications are as follows:

Answer questions about Wikipedia, answer questions about news articles, and answer questions about medical records.

Since the 1990s, great changes have taken place in the field of natural language processing technology. Two distinct features of this change are:

(1) For system input, the natural language processing system required to be developed can handle large-scale real texts, instead of dealing with very few terms and typical sentences, as in previous research systems. Only in this way can the developed system have real practical value.

(2) The output of the system is very difficult in view of the true understanding of natural language. It is not required to have a deep understanding of natural language text, but it is necessary to extract useful information from it. For example, automatic extraction of index words from natural language texts, filtering, retrieval, automatic extraction of important information, automatic summarization, and the like.

At the same time, due to the emphasis on “large scale” and the emphasis on “real texts”, the basic work of the following two aspects has also been emphasized and strengthened.

(1) Development of large-scale real corpus. The large-scale corpus of real texts processed through different depths is the basis for studying the statistical properties of natural language. Without them, the statistical method can only be passive water.

(2) The preparation of a large-scale, informative dictionary. The scale of tens of thousands, hundreds of thousands, or even hundreds of thousands of words, the richness of information (such as the collocation information of words) is very important for natural language processing.

Type1 Tethered Cable

Type1 Tethered Cable,16A Type1 Tethered Cable,32A Type1 Tethered Cable,Type 1 Tethered Charging Cable

Yangzhou JERI New Energy Co., Ltd. , https://www.jrevcharging.com