ITU researcher secures grant for personalised language learning

Teaching yourself a new language is difficult, but a new Villum Experiment awarded research project from ITU aims to make it easier. How? By tailoring the curriculum to the user’s personal interests and level of proficiency.

Mircea LunguComputer Science DepartmentResearchalgorithmsawards

Written 2 October, 2023 07:23 by Theis Duelund Jensen

As a researcher in the field of software engineering, Associate Professor Mircea Lungu from ITU’s Computer Science Department has worked at universities in many different countries. “It’s always an advantage to have a basic grasp of the language where you live, but it’s difficult to learn a new language on your own. I’ve talked to a lot of friends in similar situations about this, and that’s where the idea for a personalized language learning tool came from,” says Mircea Lungu who has just received a Villum Experiment grant to the tune of 1,99 million kr. for his project Hyper-Personalized Learning Paths Through the Internet: A Sustainable Approach To Accelerating Language Education.

The goal of the project is to develop a tool, like a browser-plugin, that acts as a written content aggregator curating a curriculum of articles, webpages, and other text material written in the language the user wants to learn. What sets Mircea Lungu’s planned language learning tool apart is its ability to aggregate materials about topics which the user has expressed an interest in. The goal is also to make the language learning tool capable of adjusting the level of difficulty to correlate with the user’s level of proficiency in the language.

 “Say for instance you want to learn French, and you are very interested in rugby – you read a lot about it online. Configuring the tool based on your personal interests will allow for a much more engaging learning experience, as the program will find French rugby content that is matched to your level of language proficiency,” says Mircea Lungu.

The system will index the foreign language Internet in such a way as to constantly present the learner with personally interesting and challenging content, and in this way, forge a hyper-personalized path to fluency for every individual learner. Since the learning content does not need to be created but rather it is discovered, this novel method can deliver personalized learning in a sustainable manner to language learners worldwide.

“The basic premise of the project is if materials that the learner would have consumed anyway in their own language can be continuously recommended in the target language at the right difficulty level, the learner will be motivated to develop regular study habits and have their path to fluency shortened.”

According to the researcher, many popular language learning applications are designed to adapt to the learner's linguistic progress but are still based on predefined generic content. They rely on gamification to keep learners engaged.

“My hypothesis is that generic content and gamification cannot compete with genuine interest. If the materials can be recommended in what in educational psychology is called the zone of proximal development – not too easy to be boring, and not too hard to be demoralizing – the learner may experience an optimized path to mastery,” says Mircea Lungu.

With funding from Villum Experiment the project has a two-year timeframe. Initial development will focus on three languages: Danish, German, and French.

Further information