Links to:

Tutorials

Four tutorials will be presented at AIED 2009:

Educational Data Mining (EDM)

Tuesday 7th July (full-day)

  • Ryan S.J.d. Baker
  • Kalina Yacef
  • Joseph E. Beck
  • Kenneth R. Koedinger

Overview

Educational Data Mining (EDM) is a key emerging area within AIED, as shown by the emergence of a conference series (held this year immediately before AIED in Spain) from a series of workshops at the AIED conference and related conferences, the emergence of a journal in the area, and the presence of large numbers of EDM papers at AIED each year. However, EDM methods are still not available to all researchers in the AIED community. This tutorial will give an introduction to key EDM methods and their concrete application within realworld data from the usage of AIED systems, so that EDM methods can be applied when useful by a broader spectrum of AIED researchers. Examples will be given from widely known research projects, and in many cases using data from the PSLC DataShop, the world's leading repository for data on the interaction between students and educational software. Attendees will learn how to access data from the PSLC DataShop for their own research, and how to make their own data available to others for future analyses within the PSLC DataShop.

Participants will come out of this tutorial with:

  • understanding of the broad areas of EDM methods
  • understanding of how these methods can be useful for studying important research topics in AIED and for improving AIED systems
  • basic knowledge of how to use popular EDM methods within their research, such that canonical versions of these methods can be directly applied
  • understanding of where in the research literature to look in order to learn more about more advanced versions of key EDM methods
  • knowledge of common pitfalls in applying EDM techniques.

Presenters

Ryan S.J.d. Baker is a Post-Doctoral Fellow at Carnegie Mellon University and Technical Director of the PSLC DataShop, the world's leading repository for data on the interaction between students and educational software. He develops and uses methods for mining the data that comes out of the interactions between students and educational software, in order to better understand how students respond to educational software, and how these responses impact their learning. He is a leading expert in automated discovery using student models, and in developing automated detectors that make inferences in real-time about students' motivational and metacognitive behavior. He was program chair (with Joseph E. Beck) of the first international conference on Educational Data Mining, and is associate editor of the Journal of Educational Data Mining.

Kalina Yacef is a Senior Lecturer at the School of Information Technologies, University of Sydney, Australia. Her research focuses on mining traces of learner's activity with educational software in order to extract patterns that can be used as an additional source of information for teaching as well as for enhancing the intelligent teaching system with predictive warnings. She is the editor (with Ryan Baker, Joe Beck and Tiffany Barnes) of the Journal of Educational Data Mining and has co-chaired 2 workshops on Educational Data Mining.

Joseph E. Beck is a Research Scientist at Worcester Polytechnic Institute. His area of research is in modeling learning within intelligent tutoring systems in order to better understand what factors influence student learning and by what amount. He has experience with constructing novel learning curve and Bayesian approaches to perform these analyses. Joe initiated the workshop series on educational data mining, eventually leading to the International Conference on Educational Data Mining, where he was program chair (with Ryan Baker) of the first conference in 2008. He is also an associate editor of the Journal of Educational Data Mining.

Kenneth R. Koedinger is a Professor of Human-Computer Interaction and Psychology at Carnegie Mellon University. He has a MS in Computer Science (University of Wisconsin, 1986) and a PhD in Psychology (CMU, 1990). He has authored over 190 papers and has won over 16 major grants. He is a co-founder of Carnegie Learning, a company marketing advanced educational technology, and is the Director of the Pittsburgh Science of Learning Center (see LearnLab.org).

Back to the top

From planning to classroom in less than 2 hours: Developing Adaptive Tutorials using LAMS and AeLP (LAMS)

Monday July 6th (full-day)

To learn more, visit the lams@aied2009 tutorial web page at: http://www.adaptiveelearning.com/news/48-aied2009tutorial

  • Dror Ben-Naim
  • Ernie Ghiglione

Overview

LAMS is a web based system for in-class activity management that enabled educators to sequence educational activities, and then deploy them to students. The AeLP is an advanced authoring environment for Adaptive Tutorials. LAMS modular architecture coupled with the "embed-ability" of the AeLP enabled the creation of powerful web 2.0 mash-up - teachers can easily embed Adaptive Tutorials in LAMS' in-class activities.
In this tutorial we will focus on the theoretical and practical aspects of developing an integrated lesson plan using LAMS and AeLP. The entire pedagogical process will be discussed:

  • Conceptualization
  • Development
  • Deployment and in-class monitoring
  • Post activity reflection and adaptation

AIED participants will:

  • learn pedagogical design principles of for intelligent content development
  • introduced to the benefits of reusable learning designs
  • create powerful adaptive learning designs
Developers and architects of educational system will also learn about web 2.0 mash-up considerations that enabled both systems to be integrated.

Presenters

Dr Ben-Naim is currently doing his PhD in the School of Computer Science and Engineering at the University Of New South Wales, Sydney, Australia. His PhD thesis titled "The Adaptive eLearning Platform" has been successfully expanded into a cross-UNSW project involving numerous Schools and Faculties. He has six years experience in developing educational software and eLearning content and has mentored students through their Honours and bachelors degrees. His research focuses on Virtual Apparatus Framework Approach to Authoring Intelligent Tutoring Systems.

Ernie Ghigloione is the LAMS Project Manager, based at the Macquarie E-Learning Centre of Excellence (MELCOE), Macquarie University. He has previous experience in various open source projects in e-learning. He has developed parts of the .LRN Learning Management System, specially the Learning Object Repository, content delivery platform, one of its assessment engines, the IMS Content Packaging, IMS Metadata and SCORM implementation. Prior to managing e-learning projects, Ernie led large enterprise software development in the US, the Netherlands and India for five years. He holds an MSc BSc Management Information Systems (magna cum laude) from New York University and a Master of Software Engineering from the University of Sydney.

Back to the top

Categorisation and Educational Benefits of Open Learner Models (OLM)

Tuesday July 7th, morning (half-day)

  • Susan Bull
  • Judy Kay

Overview

Open learner models are learner models that are directly accessible to the learner they represent (and sometimes also to other interested parties). There are a variety of reasons that a learner model might be open to the user, for example:

  • To prompt learner reflection on their knowledge and on their learning processes
  • To facilitate planning and self-monitoring of learning
  • To promote learner independence
  • To increase the accuracy of the learner model (by allowing the learner to contribute to, or negotiate the model data)
  • To afford the learner greater control over the choices in their learning o For assessment (formative and/or summative)
  • To facilitate navigation through learning materials, tasks, activities, exercises, etc.
  • To increase learner trust in an adaptive learning environment (by allowing the user to see the system's inferences about their knowledge)
  • To prompt collaboration and/or competition amongst learners (e.g. when students choose to release their learner models to peers)
  • Accommodating the learner's right to view data held about them

The tutorial will help attendees to identify areas that have not yet (or only superficially) been explored, and discuss them: i.e. produce an immediate and longer term research agenda for open learner modelling work.
We will aim to run a follow-up workshop at AIED 2011, where the ideas generated at the tutorial can be explored further, after participants have built and performed initial evaluations of novel open learner modelling approaches . Between AIED 2009 and AIED 2011, knowledge of the "SMILI Open Learner Modelling Framework" (see below) should generate more systematic descriptions and analyses of open learner models and their results, to facilitate the development of research in this area.
Immediate results of the tutorial will include a set of open learner model categorisations that will be made available to the research community to support further research, from a link on the Learner Modelling for Reflection (LeMoRe) website.

Presenters

Susan Bull is at the University of Birmingham, UK. She has given invited short research courses on open learner modelling at the International Summer School on Educational Adaptive Hypermedia 2006 and at International Interfaces for Intelligent Computer Assisted Language Learning 2006; and an invited keynote speech on open learner models at the 2004 Hellenic Conference on ICT in Education. She was coeditor of the special issue of the International Journal of Artificial Intelligence in Education on open learner models (IJAIED 17(2) and IJAIED 17(3), 2007), which brought together the first collection of papers on open learner modelling.

Judy Kay is based at the University of Sydney, Australia. She has given invited keynote speeches at many international conferences, and specifically focussing on open learner models at the 1997 International Conference on Computers in Education, and at Adaptive Hypermedia 2006. She also has a contribution related to open learner modelling in the tenth anniversary issue of the User Models and User-Adaptive Interaction journal, where papers were invited, to cover key areas in user modelling.

The University of Birmingham and the University of Sydney are the two main centres of research on open learner models. Both university departments have deployed open learner models in a selection of courses , where there has been extensive use, and user feedback.

Back to the top

Educational Natural Language Processing (ENLP)

Tuesday July 7th, afternoon (half-day)

  • Iryna Gurevych
  • Delphine Bernhard
  • Aljoscha Burchardt

Overview

Typical Web 2.0 tools such as wikis, blogs, and podcasts have recently entered the classroom and foster interactions between learners and tutors, within the new eLearning 2.0 paradigm. As a result, eLearning 2.0 makes large amounts of eLearning discourse available for Natural Language Processing (NLP) within the field of research that we call "Educational Natural Language Processing" (e-NLP). Research on e-NLP has existed for a long time and has focused on e.g. intelligent tutoring systems (Litman & Forbes-Riley, 2006), or essay scoring (Attali & Burstein, 2006). This field of research brings together two communities: language technology on the one side and educational computing on the other side. Several workshops on "Building Educational Applications Using NLP" and related topics have already taken place at major conferences, such as HLT-NAACL 2003, COLING 2004, ACL 2005, ACL 2008 and NAACL-HLT 2009.
NLP techniques are used in many educational applications working with textual data such as intelligent tutoring systems or computer-assisted language learning. However, these applications are particularly challenging for NLP since they require an adaptation of NLP techniques to various types of discourse, e.g. tutoring dialogues, which are different from typical task-oriented spoken dialogue systems. Moreover, educational applications place strong requirements on NLP systems, which have to be robust yet accurate. Therefore, this is an important application domain and a source of innovation for both NLP and educational computing, as shown by Feng et al. (2006), Kim et al. (2006), Malioutov & Barzilay (2006) and Csomai & Mihalcea (2007), to name just a few.
In this tutorial, we will review a variety of uses of NLP in the educational domain and point to emerging trends which call for new types of applications.

Outline:

  • Introduction: eLearning and NLP
  • Reading and writing assistance: readability, simplification and vocabulary assistance
  • Automatic generation of exercises for computer-based testing
  • Assessment of learner generated discourse: essay scoring, plagiarism detection, short answer assessment, speech assessment
  • Web 2.0 and computer supported collaborative learning, quality of student-generated content
  • Example e-NLP application 1: electronic career guidance
  • Example e-NLP application 2: educational question answering

Presenters

Iryna Gurevych is head of the Ubiquitous Knowledge Processing (UKP) Lab at the University of Darmstadt. Her recent research has focused on the application of lexical semantic knowledge in such areas as spoken dialogue summarization, information retrieval, and question answering for educational purposes, e.g. electronic career guidance, or question answering based on question-answer repositories in Web 2.0 applied to eLearning. Her areas of expertise include algorithms for computational lexical semantics and processing of user generated discourse. She guided the development of the high-performance Java-based Wikipedia and Wiktionary APIs as well as projects in collaborative annotation, information filtering and sentiment analysis for eLearning.

Delphine Bernhard is Senior Researcher in the Ubiquitous Knowledge Processing (UKP) Lab at the University of Darmstadt. She obtained her PhD in 2006 from the Université de Grenoble 1, where she worked on terminology extraction from domain specific texts and unsupervised morphological analysis. Her current work focuses on enhancing question answering systems to meet the specific needs of learners. Her further research topics include processing user generated discourse and quality assessment of social media content.

Aljoscha Burchardt is scientific coordinator of the Center of Research Excellence "eLearning 2.0" and Senior Researcher in the Ubiquitous Knowledge Processing Lab at the University of Darmstadt. He obtained his PhD from Saarland University in 2008, where he worked in projects related to both eLearning and applied lexical semantics. His current work focuses on the use of summarization techniques to access and present multimodal learning materials in collaborative settings.

Back to the top

Call for Tutorials (Closed)

Prospective tutorial instructors are invited to submit proposal for AIED 2009. Similar to previous AIED conferences, tutorials will run in parallel with other conference activities, such as workshops,  for either a half day or a full day. Proposals on advanced topics, new developments with a certain level of maturity and fundamentals of the AIED research are especially welcome. Tutorials should take a broad perspective on the topic, and should not be a sales pitch for a single body of work. Topics should have a direct relevance to the topics of the conference. Attendees are expected to register to the main AIED conference. 

Format of tutorial proposals

Tutorial proposals should be submitted to the tutorial chairs via email by the date specified below, should be between 2-5 pages and specify the following:

  • Title of tutorial
  • Instructor(s): name
  • Brief description of each instructor’s experience and background
  • Desired duration of tutorial: full/ half day
  • Organisation: Description of the tutorial topic and learning objectives as well as its significance for the AIED community. This may include preliminary and/or follow-up activities planned, materials to be provided to participants, etc. as applicable.
  • The intended audience (experience level and prerequisites; include list of names of potential attendees if you have a particular community in mind).
  • Expected results

Important dates

Submission of tutorial proposals: January 15, 2009
Notification of acceptance: February 20, 2009
Tutorial date: July 6/7, 2009

How to submit

Submit proposals electronically to Beatriz Barros (bbarros@lcc.uma.es) and Stephan Weibelzahl (sweibelzahl@ncirl.ie). A tutorial proposal must be a single file email attachment, either in PDF or MS Word format.


Download the Call for Tutorials PDF