Natural Language Processing (NLP) is one of the classic fields of artificial intelligence that deals with the natural language and communication side of software and robots. Since the 1950s, work in NLP has focused on building systems that allow people to communicate with and through machines in natural language. This requires basic systems for understanding language(s), being able to convert language in structured information, and vice versa: to generate narrative text from data.
Well-known NLP applications include human-machine interaction (e.g. chatbots), machine translation, text mining, and more mundane tasks such as spelling correction and writing assistance. NLP is closely connected to the sister fields of speech technology and information retrieval. NLP uses machine learning for fast deployment; the field offers rich libraries (both open source and commercial) of ready-made NLP components. Still, building NLP applications is not trivial. If an NLP tool works for English, it does not automatically work for Dutch or other languages. NLP tools can be very accurate within a specific domain, but this requires adaptation and integration of machine learning with domain expert knowledge. There is no “big green NLP button”. Also, universities produce fewer experts who know how to build NLP systems than there is demand. NLP is red hot, but how can you get it to work for you?

Antal van den Bosch is Professor of Language and Speech Technology in the Faculty of Arts at Radboud University Nijmegen, and director of the Meertens Institute of the Royal Netherlands Academy for Arts and Sciences, Amsterdam. You can read more about him on his website.