Natural Language Processing summary.

It is incredibly amazing how the human mind can easily acquire communication skills via processing language. Such a skill seemed so simple to mankind given that it requires a baby to acquire within an average of two years. However when we closely began to look at the processes that seems so natural to us, we could not help but be bewildered the complexity of communicating via words. In fact, the process is so complex that even computers, equipped with cutting edge technology and computing power developed over decades, are a long way from achieving perfection in. The studies for unravelling the secrets of communication gave birth to one of the most heavily research field in Computer Science, Linguistics etc- Natural Language Processing with the sole purpose of developing enhanced interactions between computers and natural human languages.

With the heavy usage of computers and increased reliance of the Internet, we now communicate vastly using our devices. As discussed before in cloud computing, this generates a huge amount of data that reveals (to some extent) the patterns of human communication. Moreover, with the advent of social media, we have now allocated a huge portion of our communication to be performed via internet which leads to further understanding of the human communication via language. This answers some of the most important natures of language and its usage: ranging from calculating have been recurrences of the same word in a sentence to how a particular tweet can be analyzed as happy or sad. Developing on these records, algorithms can be generated to predict the possible sentence structures that a word may belong to. Combining the rules of grammar and the knowledge of how many times the word has been used so far in a particular sentence, there are now different kind of programs that can replicate the human speech. An interesting example is a chat bot and many of them were close to passing the Turing Test. Statistical models using the probabilities of the appearance of the targeted word in a particular sentence play a key role in such mechanism of text construction.

There are also other aims of NLP. The most common ones are:

Automatic summarization: Given any text, a computer can be able to summarize it.

Machine translation: As Google Translate is growing popular by the day, we are increasingly becoming dependent on computers to translate a different language to a preferred one.

Named entity recognition: Given any text, figure out the proper nouns even if they are in different languages.

Optical character recognition: Deciphering the characters in a printed text.

Speech recognition: Given an audio clip, a computer can be able to process the information it conveys accurately.

All of them have enough room for development given the complexity of the nature of these tasks to the fundamental level. Thus this field of Computer Science has a room for research and prospective developments.

Questions:

Is there a chance in future that we can program a computer just by speaking to it with gestures and other naturals means of communicating?

What is the role of AI in NLP?

Are the statistical models made based on Bayesian probability?

References:

http://en.wikipedia.org/wiki/Natural_language_processing

http://cliqology.com/2011/04/interested-in-natural-language-processing-nlp-watch-this-amazing-ted-talk-from-deb-roy/

http://www.columbia.edu/itc/hs/medinfo/g6080/misc/articles/spyns.pdf

https://www.coursera.org/course/nlp

http://research.google.com/pubs/NaturalLanguageProcessing.html