Natural language processing has been around for a long time, but it was only after the appearance of computers that it grew out of its original field, linguistics. Nowadays, nlp is used in many areas of computer science: from word processing and search engines to virtual assistants, like Amazon Alexa.

Natural language processing can also be applied in other fields. Some months ago, I attended a fascinating Meetup event about the application of Natural Language Processing on topics concerning society and politics. The speakers, Krista Westphal and Christian Göbel, explained how they developed methods for analyzing whether a screenplay or novel passes the Bechdel test and the messages posted on Weibo to study citizen-government interactions and their political impact in China.

Yesterday I attended another event related to nlp, as interesting as the previous one, a neuroscience Meetup suggestively entitled “language: machines vs biology”. The event description is as intriguing as the title: “Over thousands of years, our language evolved from simple natural sounds to a complex system of communication, enabling us to express imaginative, conceptual ideas and creating artistic masterpieces of literature. It’s a phenomenon that only our recent technological advancements enabled us to find long-sought answers to questions such as how we process, store, and re-use sensory inputs in the brain to communicate with each other. In fact, not only with each other, but also with computers since the informational revolution. Have you ever wondered how Alexa, Amazon’s virtual assistant can understand you and carry out your commands?

To know more about these seemingly similar processes, the biology of learning and using language versus communication with computers using human languages, we invited Elisabeth Dokalik-Jonak and Tristan Miller. Elisabeth is a linguist with neuroscience background, who developed a tool for stroke patients to help language rehabilitation. Tristan is a computer scientist, specialised in natural language processing and computational linguistics, working at the Austrian Research Institute for Artificial Intelligence. Finally, Ursula Lavrenčič, co-founder of KOBI will give a pitch about their iOS/Android app that helps children who struggle with learning to read.” (Source: https://www.eventbrite.com/e/brainstorms-11-language-machines-vs-biology-tickets-65262146838)

I absolutely loved the event. As usual, I learned a lot of things and was very inspired by the presenters. These are some of the takeaways:

* The brain is learning until you die

* The key feature for learning a language is the sense of touch

* Mobile phones kill communication and language

In case you wonder how Alexa can understand humans…click here (spoiler warning: The truth is, it can’t, it’s all about text strings, commands, and programming).

After attending these two events, I started thinking about how technical writing and Natural Language Processing can work together to improve consistency and accuracy. After doing some research on the Internet I found very interesting ideas on how to combine these two disciplines. Some of them are described below:

• NLP can be used to detect non-uniform language (i.e. the same idea is expressed in different ways) in technical writing. There are already some pilot researches that apply text similarity algorithms at different language levels to find inconsistencies. These projects have even developed strategies to overcome the challenges to define rules that allow machine processing distinguish true and false occurrences in a text written by humans. Different NLP resources are being applied to analyze a technical text, for example: POS (Part of Speech) tagger, Word-Net or SVM (Support Vector Machine).

• NLP can be also helpful to mine incoherent requirements in the documentation from a linguistic perspective. In technical documents with a vast amount of requirements (hundreds of pages) it is very hard to find incoherencies manually. And here is where NLP comes into play. On the basis that the document we want to analyse for incoherency has been reviewed for consistency, requirements can be identified, tagged, paired and finally compared.
• Another application field for NLP and technical documentation is the semantic algorithms for search engines. Companies like Google use algorithms to intend what you want to know even before you type it. Google is now also capable of identifying synonyms and answering questions.

Without doubt, nlp is an amazing discipline which is developing very quickly. It will spread to all communication-related fields sooner rather than later. This opens up endless opportunities for technical writers and the development of our technical skills.