Paper Info
Title
Speech-enabled card games for language learners
Abstract
This paper debuts a novel application of speech recognition to foreign language learning. We present a generic framework for developing user-customizable card games designed to aid learners in the difficult task of vocabulary acquisition. We also describe a prototype game built on this framework that, using a Mandarin speech recognizer, provides a student of Chinese with opportunities to speak vocabulary items in a meaningful context. The system dynamically loads only the necessary vocabulary for each game in an effort to maintain robust recognition performance without limiting the lexical domain. To assess the Sentence Error Rate (SER) of our prototype, we asked college-age students from various universities in the United States and beyond to participate in a Web-based user study. The three central concepts in the game were recognized with a SER of 16.02%, illustrating the feasibility of deploying this system in a university curriculum via the Internet. Finally, to ensure that our recognizer is behaving appropriately with regard to learner speech, we perform a rigorous analysis of the recognition errors to determine their underlying causes.
Year
Venue
Keywords
2008
AAAI
user-customizable card game,speech recognition,necessary vocabulary,vocabulary acquisition,vocabulary item,generic framework,recognition error,language learner,prototype game,mandarin speech recognizer,speech-enabled card game,robust recognition performance,error rate,game design,system dynamics
Field
DocType
Citations 
Foreign language learning,Computer science,Word error rate,Curriculum,Artificial intelligence,Natural language processing,Sentence,Vocabulary,Mandarin Chinese,Limiting,Machine learning,The Internet
Conference
4
PageRank 
References 
Authors
0.89
6
2
Name
Order
Citations
PageRank
Ian McGraw125324.41
Stephanie Seneff22075364.35