<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-1701">
  <Title>Integrating Natural Language Processing into E-learning -- A Case of Czech</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Contemporary projects aiming at launching learning management systems (LMS) often focus on the introduction of an existing software tool, rather than on an innovation by means of the modern information technologies. In effect, there is almost no original research directed to the complex integration of e-learning systems with the relevant IT such as assistive technologies (dialogue systems, speech recognition and synthesis ...), knowledge acquisition and knowledge management systems, etc.</Paragraph>
    <Paragraph position="1"> Among others, the current LMS do not integrate the emerging natural language processing (NLP) applications. The adopted learning management systems often do not support even the basic functionality of a language-oriented search and retrieval of learning objects.</Paragraph>
    <Paragraph position="2"> Futhermore, the present-day LMS are not directly linked to the wealth of relevant information and knowledge sources. In the case of the higher education institutions, these sources could comprise standard libraries that provide at least an electronic catalogue of their sources nowadays, local digital libraries that are usually freely available for academics from particular institution, and the access to comprehensive electronic archives or digital libraries that are provided by many publishers and other organizations on a commercial basis. Companies often neglect valuable knowledge sources too. For example, they should consider the integration of their knowledge bases in the form of recorded questions and answers from the call centers.</Paragraph>
    <Paragraph position="3"> The current e-learning systems do not exploit the potential of available high-level personalization techniques and adaptability of the form, the content and the access to the education.</Paragraph>
    <Paragraph position="4"> Most of them cannot play the role of a showcase for the modern teaching methods.</Paragraph>
    <Paragraph position="5"> This paper surveys several areas, where NLP techniques and technologies can enhance educational systems and applications. Some of them exist in the form of prototypes only and have not been applied in an end-user system yet.</Paragraph>
    <Paragraph position="6"> Others find their place in software tools that have been implemented by our team. They will be briefly introduced in the paper.</Paragraph>
    <Paragraph position="7"> The range of LMS used or tested at Masaryk University, Brno, Czech Republic (MU) is rather broad. The most important ones are IL-IAS (http://www.ilias.uni-koeln.de/) and MOODLE (http://moodle.org). Both systems are developed and distributed under the term of the GNU General Public License and provide open platform appropriate for the integration on NLP solutions. The actual project at MU aims at unification of the used e-learning platforms and their integration with the administrative information server of the university (Pavlovic et al., 2003). Even though separate systems would be more modular, easily maintainable and extendable, we opt for the integrated solution that will benefit from the permanent technical support and personal assistence of the administrative information server team. We strongly believe that NLP techniques as a part of the e-learning system can help to open doors to those faculties and departments that have not discovered the world of computer-mediated education yet.</Paragraph>
    <Paragraph position="8"> The paper discusses also the incorporation of languageresourcestosupportthelearnerduring his/her interaction with an educational system and to provide personalized learning. We also tackle the use of NLP technologies and resources to support the automatic assessment of learners' answers, especially those which are in free text or restricted free text form. Such assessment is useful to learners for controlling their learning progress (self-regulation), to teachers for gathering information about learners and to systems for personalizing interaction. Concept mapping is a knowledge elicitation technique, which stimulates learners to articulate and synthesize their actual states of knowledge during the learning process. We propose the use of NLP in concept mapping systems in order to interactively support learners, who build concept maps and automate the process of the assessment of concept maps. The availability of wordnet-like semantic networks resulting from several projects such as EuroWord-Net (Vossen, 1998), BalkaNet (BWN, 2004), RussNet (Azarova, 2004), or broad-coverage ontologies such as SUMO (SUMO, 2003) provide a reasonable starting point for such an effort.</Paragraph>
    <Paragraph position="9"> The NLP applications in the area of e-learning can be divided according to various criteria. They can be specific for synchronous or asynchronous mode of the course. The main focus of the methods can be stressed to address e.g. enhancements of the teaching material accessibility or the adaptability of LMS.</Paragraph>
    <Paragraph position="10"> Also the complexity of the needed NLP techniques can make the distinctions, whether the methods are already available and prepared to integration into LMS or they need further development. The availability of language resources or language technology (lingware) for the particular language can make the difference too. A related issue can be the portability of a solution for other languages or other subject area, where subject-specific information cannot be obtained fully automatically. One can also divide the NLP applications in e-learning according to the NLP modules that are integrated, e.g. a language-specific morphological module or named-entity analyzer could play a crucial role. As the educational process has two faces -- learning and teaching, the boarder-line can also be drawn between the tools focusing on the students' side and those intended for the course authors and teachers.</Paragraph>
    <Paragraph position="11"> The last mentioned aspect has been taken into account in this paper. It is organized as follows: The next section discusses NLP techniques aimed at enhancements for the end-users of e-learning systems -- students looking for an appropriate e-learning material or those who already enrolled in a course. The third section tackles the support of authors and providers of the e-learning facilities that can take the advantage of the language and text technology too. Of course, the boundary between those two cases is not strict at all, so there are NLP tools that can help both types of LMS users. The fourth section then covers supplementary information technologies such as multimedia and audio- or video- recording of courses that cannot be classified as NLP per se but are strongly related and, as our experience already shows, their integration should be at least coordinated with the employment language technology solutions. The paper concludes with future directions of our research.</Paragraph>
  </Section>
class="xml-element"></Paper>