8 ways you are already using linguistic computing in your daily life
“Humanities and science students often do not understand each other because there is a war between the disciplines. We have to destroy that mindset today,” said Professor Marco Passarotti, the programme director of Università Cattolica del Sacro Cuore’s new Linguistic Computing master’s degree.
Computer science can be used to help linguists better understand and analyse language. Similarly, language is used in computer science to advance programming. Linguistic computing combines the two, focusing on processing, managing and structuring large sets of linguistic data.
How are you already using linguistic computing?
Today, we have a growing collection of data online. “They say that if something does not cost, this means that you are the product. This is the case for Google, for Amazon, for any social network,” said Professor Passarotti.
“Every day, millions of words are added to the web by users on blogs, social media, emails and articles. Today the real challenge is to move from information to knowledge. How do companies make use of this data, this knowledge? This is where linguistic computing comes in.”
Linguistic computing is a large area which encompasses many disciplines. One of the most relevant of these is computational linguistics.
Here are just a few ways you might be already using linguistic computing in your daily life:
AI assistants: Siri, Alexa, Google Assistant
Computational linguistics led to the development of AI assistants such as Siri and Google Assistant. Professor Passarotti explains how our phones make use of linguistic computing on a daily basis.
“Computational linguistics is everywhere,” he said. “I have an iPhone, and if I say ‘hey Siri’ I am performing linguistic computing because a speech input is translated into text. The text is then analysed by Siri.
“Siri understands that it is a question. It searches for a question in an enormous amount of linguistic data on the web and searches for an answer. It then moves from text to speech when answering me, using my language. In my case, this is Italian.”
"Most of us use search engines daily. What you may not know is that every time you run a query on a search engine, linguistic computing is taking place", explains Professor Passarotti.
“You are writing words and Google searches for these words in an enormous collection of words on the web,” he said. Google collects this data and sorts it, producing search results which are then usable to you.
"If you do a Google search for something you would like to buy online, such as a new table", says Professor Passarotti, "the search engine will search the word in many different languages and then present the findings to you. This is an example of the application of machine translation, a form of linguistic computing".
In this same Google search, the search engine will perform morphological processing: finding all of the different word forms of the item.
For instance, the word ‘table’ has two forms: the singular ‘table’ and the plural ‘tables’. Both are searched by the engine regardless of the fact that when you typed the query you chose one form or the other.
If you have ever talked about something and then had it appear in a banner advert on a website, you will have experienced linguistic computing in action.
To test this, Professor Passarotti mentioned Tübingen, Germany in front of his television every day to see if he would start seeing adverts for it. Sure enough, after about four days, he started receiving banner advents on his Facebook account about flights to Tübingen.
He said: “I did not type anything into Google. I did not search for any hotel on Booking.com. This means that the television hears me, understands what I'm talking about and that data is structured into information, which becomes knowledge. This knowledge is sold to Facebook and Facebook gives me adverts.”
Despite the divide between the humanities and sciences, linguistic computing is frequently used in academia across both disciplines.
Professor Passarotti gives the example of processing a large number of classical texts, such as the texts of Ceasar. A linguist’s research could be optimised by using a computer program that automatically analyses all of this text.
Not only would this save the linguist time, but it would significantly increase the quality of the research that the linguist is able to do.
Similarly, linguistic computing can be used to help companies with their market research for particular products.
Professor Passarotti said: “Some years ago, I took on a consulting project for McDonald's. They wanted to know the public’s opinion on a new salad they had launched. They wanted me to perform an opinion mining task on the web.
Here I used linguistic computing to mine the web and extract the opinions that people have on the salad, which McDonald’s could then use as feedback.”
Banks and insurance companies can use linguistic computing to organise and extract information from large sets of data, such as emails.
Hospitals also use linguistic computing to sort the clinical files of patients. These files may span 40 years and are made up of many words. Mining this linguistic data is necessary to translate this information into knowledge which can be used in the hospital.
Why study a master’s in linguistic computing?
Now that you understand how linguistic computing is used in day to day life, you may be wondering what you could learn from a master’s in linguistic computing.
The new MSc Linguistic Computing degree is offered by Università Cattolica del Sacro Cuore’s Faculty of Linguistic Sciences and Foreign Literatures.
The programme was created to fill a gap in the job market for graduates who are highly skilled in humanities and computer science. “Every day I receive emails from employers searching for people with expertise in linguistic computing,” said Professor Passarotti. “Every company needs somebody who can perform automatic analysis of data.”
The master’s in Linguistic Computing will provide advanced knowledge of the problems, methods, techniques and tools related to the automatic processing of linguistic data.
It prepares professionals to analyse all forms of linguistic data in different sectors.
How is the Linguistic Computing degree taught?
"If you are looking for an easy master’s degree, a master's in Linguistic Computing at Università Cattolica is not for you", warns Professor Passarotti.
“If you want to do big things, you have to study something that seems to be outside your comfort zone,” he said. “This is not an easy degree. Students are paying for this programme and I want to give them the best experience I can. I want people who are surprised about what they see and who genuinely get excited by this.”
Learning the fundamentals of linguistic computing
The two-year programme degree focuses on theoretical teaching in the first year, moving on to the practical elements in the second year.
“During the first year, we train students in the fundamentals of computational linguistics, with different layers of linguistic and notation. We provide them with a solid background in formal methods for natural language processing,” said Professor Passarotti.
This includes learning fundamental computer skills, digital language resources and automatic language processing methods.
Students will also be introduced to project management, learning how to manage large projects in various areas of business.
Practical application of skills
The second year sees these skills put into practice through practical activities, internships and laboratory work. Students learn how to develop and apply automatic language processing tools in various humanities fields, such as libraries, archives, museums, publishing houses and social media; or using the tools for management orientated business applications.
Professor Passarotti said: “Companies, such as Google, are supporting our linguistic computing degree. Students can complete internships in these companies, which can serve as an entrance to the professional world.
“This degree is theoretical, it's practical, it's company related. Ultimately, everything deals with the computational way of looking at the linguistic data. This is to meet the inevitable turn that we are facing in today’s job market.”