Data scientist Larry Finer has developed a range of artificial intelligence and machine learning skills. One part of Dr. Larry Finer’s work in AI consists of finding uses for natural language processing (NLP) in data analysis.

Natural language processing is a simple concept that is difficult to implement. A computer using NLP processes data on the words used by human beings in everyday situations and summarizes the results. This process involves machine learning — discovering information by making inferences from data rather than relying on external programming.

An NLP-mediated exchange between a human being and a computer works like this: The computer converts the sounds of human speech from audio to text, processes the text, determines a response, then turns it back into an audio file and plays back the results. Applications for this process cover functions such as checking text for grammar, translating languages, and responding to phone calls, as well as powering software such as Alexa and Siri.

The idiosyncrasies of human language make NLP difficult. For instance, NLP has trouble detecting sarcasm and often is not attuned to ambiguities in usage. Communicating via NLP involves two distinct tasks:

– Syntactic analysis. Computers must know the rules of grammar to understand human utterances. NLP systems must divide the flow of sounds into discrete words, identify their functions within a sentence, and know when sentences stop and start.

– Semantic analysis. This is the assigning of meaning; it is one of NLP’s greatest challenges. It works by grouping speech into predetermined categories, such as names and places. NLP must also decode word meanings by their context, and compile huge databases of these definitions.