Judging Language at Scale: Assessment and Feedback Inside an AI Language-Learning Platform
18 novembre 2026
Alyx is an AI-powered language-learning platform where learners practise conversation, complete structured lessons and work on pronunciation with AI teachers, while their human teachers follow their progress.
In this webinar we open up how the system works. We’ll walk through how learner activity is evaluated, from real-time correction of each conversational turn to end-of-session assessment and pronunciation scoring, and show the feedback that learners and teachers actually receive.
We’ll also discuss what makes this hard. A rule that holds in English can misjudge a correct sentence in French or German, acceptable variation has to be told apart from real error, and a single grade has to mean something to both a learner and a teacher. The session is designed for researchers in linguistics and computing interested in automated language assessment, learner modelling and educational feedback.
This webinar is organized in the framework of the LIDILEM research action EV⋅AL and the upcoming Doctoral Networks Horizon MSCA HODACE (2027-2031).
Guest Speaker
Raíza Hanada (Research Lead, Alyx AI)
About Alyx AI
You can access Alyx AI's website here: https://alyx-ai.com/
Webinar official web page: https://lidilem.univ-grenoble-alpes.fr/actualites/webinar-judging-language-scale-assessment-and-feedback-inside-ai-language-learning-platform
Infos
- Sylvain Coulange
- 18 novembre 2026 14:48
- Conférences
- Anglais
Commentaire(s)