É«É«À²

Aalto University and the University of Helsinki are involved in a project, which has produced a learning game that can be played on your own smartphone or tablet computer anywhere and as many times as you want. Thanks to the voice-user interface, the game is suited as a learning environment also for children unable to read as well as illiterate adults, who represent seven per cent of the refugees arrived to Finland.

‘It is expensive to organise teaching, so we need technological solutions to support learning that are suited for different age groups and take also users unable to read into account. The learning game can also be used in schools, where it is difficult and expensive to organise practice in active speaking and immediate personal feedback due to large group size,’ says Professor Mikko Kurimo from Aalto University.

New games using speech recognition

The project offers companies a way of applying open source tools developed by researchers to their own projects. Gaming world companies in particular can add speech recognition technology to their games and thus open doors for totally new kind of games and other applications where speaking is the most efficient way of communicating with mobile devices.

‘Computer-assisted language learning and different learning games are already quite popular, but most of them teach mainly writing skills. Since no sufficiently advanced speech recognition technology has been available, giving feedback on oral communication has been very limited and implemented on a few languages only. The speech recognition technology being developed in this project may generate new innovations in such field as game industry or education,’ Mikko Kurimo points out.

Points for correct pronunciation

The idea of the game is to explore various game boards and to turn cards which can teach words in the target language. When you click the card, you see a picture and a voice says the word first in your native language and then in the target language. The players repeat the word in the target language, after which they can evaluate their own performance by listening to their own pronunciation and then to the model pronunciation. As feedback for pronunciation, the game's speech recognition system gives 1 to 5 stars, and collecting them takes you forward in the game. The game has 27 levels.

The Technology-Enhanced Language Learning with mobile devices (TELLme) project won funding in Tekes' Challenge Finland competition. The project involves a team headed by Mikko Kurimo at Aalto University and a team headed by Sari Ylinen at the University of Helsinki from the Cognitive Brain Research Unit, which is one of the world's leading units in brain research and application of its results in the study of learning. The Aalto University research group, on the other hand, is one of the global leaders in the field of automatic speech recognition and its challenging real-world applications.

Watch the video to see how the game is played: