Automatic Speaking Assessment Systems for Second Language Learner's Spontaneous Speech in Under-Resourced Languages

Explore an in-depth lecture on developing automatic speaking assessment systems for second language learners' spontaneous speech in under-resourced languages. Delve into the challenges and recent advancements in creating these systems for languages with fewer learners, such as Finnish and Swedish. Learn about the demo system for Finnish and Finland Swedish, its main components, and evaluation results. Discover how personalized feedback is provided to language learners and gain insights into the most important input features for machine learning models. Examine the system's multiple components, including Automatic Speech Recognition, Lexico-grammatical Accuracy and Range Evaluators, Pronunciation and Fluency Evaluators, Task Accomplishment Evaluator, and a CEFR-like score Evaluator. Gain valuable knowledge from Professor Mikko Kurimo, an expert in automatic speech recognition and spoken language modeling, as he shares his extensive experience and research in the field.