When developing modern artificial intelligence (AI) solutions, scientists tend to follow heuristics that are based on a module-based approach. For example, a team of developers working on an advanced smartphone app to help the legal industry will likely dedicate resources to work on separate modules such as machine learning, case-based reasoning, user interface, natural language processing, etc.
Of all AI development modules, natural language processing (NLP) is currently one of the most crucial. NLP can be defined as the ability of AI solutions such as computers or software applications to understand humans. At one of its most basic levels these days, NLP can be exemplified by the queries handled by search engines such as Google.
When a Google Search user types “what is the meaning of life?” versus “Meaning of Life,” the search engine algorithm will recognize that the former query could be a philosophical question due to the subject matter and the question mark at the end while the latter could be a search for information about the 1983 film by British comedic troupe Monty Python. This textual interpretation and discernment is part of NLP.
The Importance of Voice Recognition in NLP and AI Development
The NLP module of AI development is of great interest among scientists, futurists and linguists. One of the strongest arguments in favor of AI is that it should be developed to serve humanity in a smart and meaningful way. This argument can be easily construed as a need for AI solutions to communicate with humans.
Linguists and foreign language teachers can tell you with certainty that understanding is the key to learning. First language acquisition is one of the most complex intellectual achievements by humans; we may not realize this because we typically acquire our first language during the wondrous period of childhood, which means that we don’t remember how difficult it was for us to assimilate the intricacies of communications. We come to this realization when we are adults trying to learn a second language for the purpose of keeping a job.
NLP is such an essential component of AI development that social networking giant Facebook has about 100 researchers working on AI projects that focus on solutions that are skilled enough in language that they can hold basic conversations with humans.
How Voice Recognition Will Change AI
NLP is more than Apple Siri, Google Now and Microsoft Cortana playing MP3s, taking dictation or looking up search engine queries. That is voice recognition. Advanced speech recognition, on the other hand, will completely change the relationship humans have with AI systems. Speech recognition would lead to more applications such as Watson, the Jeopardy champion developed by IBM.
When Siri or Cortana are able to understand the nuances of language, they will be able to tell us more than what we would like to know. In this fashion, a Siri user who hurriedly barks the question “How far from Miami to Chicago?” should get more than just a distance in miles; an advanced Siri may detect the hurried tone of the user’s voice and provide traffic reports, information about rental car agencies, flights, road conditions, etc.
Can You Have AI Without NLP?
Since humans are developing AI solutions, it is only natural to think that NLP and speech recognition are high on the list of achievements. The development of the NLP AI module is intrinsically tied to the advancement of two other modules: Deep Learning and Artificial Neural Networks, which happen to be at the cusp of AI reasoning.
Think about a five-year old girl who is learning to hold conversations; as she becomes eloquent in the second grade of elementary school, has she only become a chatterbox? Not at all; her eloquence is but one aspect of her intellectual development, and we can safely assume that this little girl is becoming very smart. A similar development path can be expected of AI with regard to NLP.