EDUCATIONAL QUESTION ANSWERING BASED ON SOCIAL MEDIA CONTENT

Abstract

In recent years, the amount of digital textual information has been constantly increasing,
leading to the well-known information overload problem. While this problem is especially acute for
learners, conventional search engines are often ill-suited to address learners’ complex information
needs. We believe that Question Answering (QA) represents a more appropriate Natural Language
Processing (NLP) technology in educational conexts, both to reduce the learners’ information
overload and the instructors’ work overload. On the one hand, learners have to deal with a growing
amount of learning and community-based material in which to look for relevant information. On the
other hand, instructors are overwhelmed with students’ questions asked via forums or emails. These
challenges should be addressed by an educational QA system which could automatically answer a
significant part of the students’ questions. Educational QA would thus constitute a significant
technological asset for independent and technology-enhanced learning. QA systems actually share
some interesting characteristics with other learning technologies. They provide a means for learners
to obtain answers to their questions, just as forums and chats. However, QA systems are not
dependent on human responses and thus cater for timely responses. They are also related to
Intelligent Tutoring sytems, though