High functioning natural language processing is hands down one of the most sought after features of interactive voice response systems. Natural language processing (NLP) is a field that combines computer science and linguistics and studies interactions between human language and computer processing.
As the above title suggests, the nuances and slight variations included in every human language make this one of the most difficult and confusing fields in all of science. How would one teach an inanimate object that doesn’t have consciousness or inherent lingual abilities to distinguish between two words that have very different meanings but sound phonetically the same?
NLP is a sub-branch of artificial intelligence, and is primarily focused on developing technology that allows machines to interpret verbal communication that enables humans to speak in their natural language, dialect, and accent.
So where did this idea of natural language processing originate? Presumably, computers have only been functional on a consumer level for several decades. It actually dates back much further than most would assume, and can be traced to patents applications in the 1930s for a translating machine. The machines would in theory be bilingual and able to translate between languages with varying levels of complexity (including dealing with grammatical roles between languages).
Actual software capable of performing NLP tasks was not released until 1954 during an experiment at Georgetown University and IBM. Researchers were able to program an application to fully and automatically translate more than sixty Russian sentences into English.
There was a ten year gap between the development 1954 and 1964 (when an application was released that could solve algebraic word problems), put post-1964 there was a steady stream of NLP applications released. Some of the more interesting ones included chatterbots able to simulate natural human language and generate prose at random and a question answering system that was so advanced that it was able to win Jeopardy when pitted against some of the best human contestants in the world (most will recognize the device as Watson).
When NLP programs are integrated into software applications and IVR systems, the results are lauded as the type of programs that science fiction would imagine. The actual technology behind NLP applications is extremely complex and requires extensive research and testing to implement.
NLP software and applications have been developed that perform a variety of language interpretation tasks. Various tasks include translating text from one human language to another, identifying the discourse structure of text, converting information stored in databases to readable language, offering answers to human-language questions, and even identifying the sentiment of spoken or written text based on extracting subjective information.
Coming up: The challenges and complexities of programming a machine to understand a person.
6