What is natural language processing and what is it used for?
Natural language processing (NLP) is the processing of natural language information by a computer. It is a subfield of computer science closely associated with artificial intelligence, and its applications include speech recognition, machine translation, sentiment analysis, grammatical error correction, and analysis of electronic health records.
When did natural language processing begin and who started it?
NLP has its roots in the 1950s. In 1950, Alan Turing published "Computing Machinery and Intelligence," which proposed the Turing test and included automated interpretation and generation of natural language as a criterion of machine intelligence. The Georgetown experiment in 1954 demonstrated fully automatic translation of more than sixty Russian sentences into English.
What was the ALPAC report and how did it affect NLP research?
The ALPAC report was published in 1966 and concluded that ten years of machine translation research had failed to meet expectations. Its findings caused funding for machine translation to be dramatically reduced, and little further research in machine translation was conducted in America until the late 1980s.
What is the difference between symbolic NLP and statistical NLP?
Symbolic NLP relies on hand-coded rules and dictionary lookups to process language, while statistical NLP uses machine learning algorithms trained on large bodies of text. Statistical methods, introduced widely in the late 1980s, are more robust to unfamiliar and erroneous input and scale in accuracy as the amount of training data grows.
What role did Word2vec play in natural language processing?
Word2vec was developed by Tomáš Mikolov, who began his work as a PhD student at Brno University of Technology around 2010 by applying a simple recurrent neural network to language modeling. Word2vec helped establish representation learning and deep neural network methods as the dominant approach in NLP through the 2010s.
What was the first machine-generated book and when was it published?
The first machine-generated book was "The policeman's beard is half-constructed" by Racter, created by a rule-based system in 1984. The first published work by a neural network was "1 the Road" in 2018, a novel containing sixty million words. The first machine-generated science book, "Lithium-Ion Batteries" by Beta Writer, was published by Springer in 2019.