Abstract:

The ICE3 project uses state-of-the-art learning platforms to provide automatic feedback to language learners. Other less sophisticated learning platforms provide automatic feedback and work with simple pattern matching processes. These test whether, for example, a string a student puts in to a gap-fill exercise is exactly the one expected or ,in the case of a multiple choice exercise, whether the student has made the right choice from a number of options. A typical example the kind of feedback given for the gap fill exercises would ‘The first two letters of your answer are correct.’ Feedback of this kind is pedagogically demotivating.

The Autolearn platform used in this ICE3 project was developed from an earlier project (ALLES) to extend the use and application of intelligent feedback in language learning (autolearn.barcelonamedia.org).

AutoLearn is currently available for learning English or German as a foreign language with the help of Automatic Correction facilities which go beyond mere spell and grammar checking. It includes a facility for checking the content of an answer the checking of correctness in terms of content. It checks whether the student’s input is a factually and grammatically acceptable solution to the tasks set. This extended functionality is based on advanced intelligent natural language processing (NLP).
AutoLearn allows learners to get instant feedback from the tutoring system but can also be combined with teacher supervision. It uses the Moodle open-source content-management system which means that many of the Moodle collaborative learning functionalities can be also exploited for language learning.
In order to improve usability of these correction tools for language teaching, an additional tool was developed that allows language teachers without computational linguistic skills to provide automatic correction exercises for learners.
AutoLearn uses advanced NLP techniques in combination with the much less sophisticated ‘Hot Potatoes’ exercises. The goal is to use computational devices to analyse learners’ production in order to go beyond “yes-or-no” answers. Intelligent feedback, as provided by Autolearn, will increase learner awareness of appropriate language and communication strategies.
AutoLearn successfully tested the tools that it developed with a number of class groups in university, high school and further education environments. It demonstrated that the use of automated feedback in response to second language learner production is feasible and beneficial for both learners and teachers in an appropriate pedagogical ICT setting. The AutoLearn tools will now be further tested and extended in this new project: ICE3 (ice3.barcelonamedia.org). The main goals of ICE3 are the following:

- To evaluate CALL-based activities in school contexts from the perspective of developers, teachers and learners alike.
- To involve teachers in the material creation process, exploiting the possibilities of ICT, generally within the context of the communicative approach, thus enhancing learners’ autonomy and teachers’ creativity.
- To incorporate the Spanish language into the platform with spell, grammar and content checking
AutoLearn tools exploit the potential that Language Technologies have to analyse responses written by second language learners. They are currently being extensively tested to ensure that they can be offered to the education community at large.

Keywords:

CALL, educational technology, language learning, language teaching, natural language processing.