Information Retrieval for Reading Tutors

Reading tutors are one example of an intelligent tutoring system. Such intelligent tutors seek to integrate content knowledge and instructional principles with models of an individual learner's knowledge and beliefs. In this way, an intelligent tutor can provide individualized instruction appropriate to the stage of development of the learner in the targeted knowledge domain. In this context, the use of information retrieval (IR) for reading tutors is relatively recent (Brown & Eskenazi, 2004) and stems from the strong desire on the part of language educators to give students practice in a realistic environment that goes beyond second language textbooks. In this case, the realism is provided by the authentic texts that are retrieved from the Internet. An authentic text is a text that has not been created or edited specifically for a (second) language learner. The challenge in using authentic documents to teach reading and vocabulary is that, unlike documents prepared by teachers expressly for their students, the materials typically obtained from the Internet are not homogeneous in content and in difficulty. This means that a simple search, containing one or several words and using a common search engine such as Google, will not retrieve documents that are satisfactory for use in tutoring. To find usable documents, the content of the Web pages that are found must be filtered, annotated, and presented to the student in an appropriate environment. Keywords: CALL; esl/efl; language learning technology; reading