Corpora Processing and Computational Scaffolding for a Web-based English Learning Environment

Abstract

This paper describes the development of an innovative web-based environment for English language learning with advanced data-driven and statistical approaches. The project uses various corpora, including a Chinese-English parallel corpus (Sinorama) and various natural language processing (NLP) tools to construct effective English learning tasks for college learners with adaptive computational scaffolding. It integrates the expertise of a group of researchers in four areas: (a) advances in NLP technologies and applications, (b) construction of a self-access reading environment, (c) exploration of English language learning through written exercises and translations, and (d) use of bilingual corpora for culture-based English learning. In this paper, the conceptualization of the system and its various reference tools (e.g., a bilingual concordancer) for English learning and pilot testing on various modules (e.g., a reading module, Text Grader, and Collocation Practice) are reported.