Ebook: Artificial Intelligence in Education

The 14th International Conference on Artificial Intelligence in Education (AIED2009) is being held July 6–10 2009 in Brighton, UK. AIED2009 is part of an ongoing series of biennial international conferences for top quality research in intelligent systems and cognitive science for educational computing applications. The conference provides opportunities for the cross-fertilization of techniques from many fields that make up this interdisciplinary research area, including: artificial intelligence, computer science, cognitive and learning sciences, education, educational technology, psychology, philosophy, sociology, anthropology, linguistics, and the many domain-specific areas for which AIED systems have been designed and evaluated.

AIED2009 focuses on the theme “Building Learning Systems that Care: From Knowledge Representation to Affective Modelling”. This extends an AIED vision proposed 20 years ago by John Self. The field has moved a long way since then. It is now widely accepted that effective learning environments are expected to care about both learners and tutors, and to have a good understanding of the variety of learning contexts. The key research question now is how to tackle the complex issues related to building learning systems that care, ranging from representing knowledge and context to modelling social, cognitive, metacognitive, and affective dimensions. This requires multidisciplinary research that links theory and technology from artificial intelligence, cognitive science, and computer science with theory and practice from education and the social sciences.

AIED research is widely recognized and has been in the focus of recent funding opportunities and initiatives all over the world. One of the goals of the AIED2009 organizers was to encourage broad participation from research communities, users, and educational practitioners whose work is relevant to AIED. The conference attracted submissions from 39 countries all over the world, including 19 European countries, 13 Asian countries, 4 American countries, 2 countries from Australia and Oceania, and 1 African country. Most submissions came from North America (45%) and Europe (37%), and there has been a noticeable increase in participation from Asia and Australia and Oceania (15%). Overall, we received 243 submissions of full papers and posters; 70 of these were accepted as full papers and another 70 as posters. The full papers have been allotted 8 pages in the Proceedings whereas the posters have been allotted 3 pages. The conference includes a Young Researchers Track combined with a Doctoral Consortium (YRT/DC). This is the main forum for PhD students and new researchers to present their work as primary investigators and to receive feedback from the international AIED community. AIED2009 received 25 YRT/DC submissions, 18 of which were accepted for presentation at the conference and were allotted 2 pages in the Proceedings. The YRT/DC session is supported by the US National Science Foundation.

In addition to the main research tracks, the conference includes 12 events that showcase interactive demonstrations of AIED systems. There also are 12 workshops held in conjunction with AIED2009 that provide an opportunity for in-depth discussion of current and emerging topics of interest to the AIED community. In parallel to the workshops, 4 tutorials present advanced topics and current developments that have a level of maturity in AIED research.

The invited speakers present key aspects of AIED09 theme and point to prospective AIED research directions. Susanne Lajoie revisits the traditional AIED debate as to whether computers can teach you to think and care, looking at affect, motivation and meta-cognition. Coming from AI and cognitive science, Kenneth Forbus places his work on the understanding of open-domain sketches in the context of AIED. Another driver for future AIED research, namely the recent developments in the world wide web, is presented by Wolfgang Nejdl, who discusses how user generated content can be used to improve search.

The AIED2009 program committees (PCs) played a crucial role in shaping the conference programme. Senior PC members reviewed YRT/DC submissions and provided constructive feedback and guidance to the next generation AIED researchers. Each paper and poster submission was reviewed by three members of the program committee, with a member of the senior PC overseeing the review process and providing a meta-review based on the reviewers' scores for novelty, originality, soundness, and relevance to the conference. In some cases, the PC members nominated external reviewers which encouraged many establishing AIED researchers to take part in the review process and provide invaluable feedback about the quality of submissions. The review process was rigorous; many submissions about promising research had to be rejected due to the healthy competition.

A conference of this size could have not been organised without the active involvement and invaluable contribution of the organizing committee. We are indebted to the YRT/DC chairs George Magoulas and Tanja Mitrovic, Poster chairs Neil Heffernan and Tsukasa Hirashima, Interactive events chairs Jack Mostow and Katy Howland, Workshop chairs Scotty Craig and Darina Dicheva, Tutorial chairs Beatriz Barros and Stephan Weibelzahl, and Sponsorship chairs Roger Azevedo and Rose Luckin. Without them, this conference could have not happened. Special thanks go to the Publicity chair Genaro Rebolledo Mendez for his enthusiasm and creativity in popularising the conference to the relevant research communities Many thanks are also due to the local team at Brighton, especially Katy Howland, Madeline Balaam, Amanda Harris and Hilary Smith. We are also grateful for all the help received from Julia Gallagher and Darren Johnsonat VisitBrighton. The AIED2009 conference promises to be a stimulating research event, presenting the state-of-the-art projects and shaping the future of AIED research. We encourage you join the international AIED society which brings together a community of researchers through the organization of Conferences, the International Journal of Artificial Intelligence in Education, and other activities of interest.

Vania Dimitrova, University of Leeds, UK

Riichiro Mizoguchi, Osaka University, Japan

Benedict du Boulay, University of Sussex, UK

Art Graesser, University of Memphis, USA

May 2009

More and more information is available on the Web, and the current search engines do a great job to make it accessible. Yet, optimizing for a large number of users, they usually provide good answers only to “most of us”, and have yet to provide satisfying mechanisms to search for audiovisual content. In this talk I will present some ongoing work at L3S addressing these challenges, done in the context of several European Union funded projects on personal information management and web search.

Regarding personalization, I will talk about personalizing Web Search based on user content, which goes beyond simple user profiles used in other systems. The algorithms presented improve Web queries by expanding them with terms collected from each user's personal information repository, thus implicitly personalizing the search output. Generating the additional query keywords is done by analyzing user data at increasing granularity levels, ranging from term and compound level analysis up to global co-occurrence statistics, as well as to using external thesauri. Extensive empirical analysis shows some of these approaches to perform very well, especially on ambiguous queries, producing a very strong increase in the quality of the output rankings.

Regarding search for audiovisual content, I will focus on exploiting user generated information, and discuss what kinds of tags are used for different resources and how they can help for search. Collaborative tagging has become an increasingly popular means for sharing and organizing Web resources, leading to a huge amount of user generated metadata. These tags represent different aspects of the resources they describe and it is not obvious whether and how these tags or subsets of them can be used for search. I will present an in-depth study of tagging behavior for different kinds of resources - Web pages, music, and images. The results are promising and provide more insight into both the use of different kinds of tags for improving search and possible extensions of tagging systems to support the creation of potentially search-relevant tags.

We explored the possibility of predicting learners' affective states (boredom, flow/engagement, confusion, and frustration) by monitoring variations in the cohesiveness of tutorial dialogues during interactions with AutoTutor, an intelligent tutoring system with conversational dialogues. Multiple measures of cohesion (e.g., pronouns, connectives, semantic overlap, causal cohesion, coreference) were automatically computed using the Coh-Metrix facility for analyzing discourse and language characteristics of text. Cohesion measures in multiple regression models predicted the proportional occurrence of each affective state, yielding medium to large effect sizes. The incidence of negations, pronoun referential cohesion, causal cohesion, and co-reference cohesion were the most diagnostic predictors of the affective states. We discuss the generalizability of our findings to other domains and tutoring systems, as well as the possibility of constructing real-time, cohesion-based affect detectors.

This paper describes the use of sensors in intelligent tutors to detect students' affective states and to embed emotional support. Using four sensors in two classroom experiments the tutor dynamically collected data streams of physiological activity and students' self-reports of emotions. Evidence indicates that state-based fluctuating student emotions are related to larger, longer-term affective variables such as self-concept in mathematics. Students produced self-reports of emotions and models were created to automatically infer these emotions from physiological data from the sensors. Summaries of student physiological activity, in particular data streams from facial detection software, helped to predict more than 60% of the variance of students emotional states, which is much better than predicting emotions from other contextual variables from the tutor, when these sensors are absent. This research also provides evidence that by modifying the “context” of the tutoring system we may well be able to optimize students' emotion reports and in turn improve math attitudes.

Affect has been the subject of increasing attention in cognitive accounts of learning. Many intelligent tutoring systems now seek to adapt pedagogy to student affective and motivational processes in an effort to increase the effectiveness of tutorial interaction and improve learning outcomes. However, the majority of research on tutorial feedback has focused on pedagogical content, often at the expense of the affective component of the learning process. It is unclear under which circumstances it is more appropriate to focus directly on student affect and when support is best offered through task-related feedback. This paper proposes an inductive framework for modeling task-based and affect-based feedback to inform the behavior of pedagogical agents within a narrative-centered learning environment.

This study shows that affect-adaptive computer tutoring can significantly improve performance on learning efficiency and user satisfaction. We compare two different student uncertainty adaptations which were designed, implemented and evaluated in a controlled experiment using four versions of a wizarded spoken dialogue tutoring system: two adaptive systems used in two experimental conditions (basic and empirical), and two non-adaptive systems used in two control conditions (normal and random). In prior work we compared learning gains across the four systems; here we compare two other important performance metrics: learning efficiency and user satisfaction. We show that the basic adaptive system outperforms the normal (non-adaptive) and empirical (adaptive) systems in terms of learning efficiency. We also show that the empirical (adaptive) and random (non-adaptive) systems outperform the basic adaptive system in terms of user perception of tutor response quality. However, only the basic adaptive system shows a positive correlation between learning and user perception of decreased uncertainty.

We researched the impact of gendered pedagogical agents on student attitudes for math, motivation and achievement in math, within the context of an adaptive tutoring software for high school mathematics. Learning companions emphasize perseverance by valuing effort in challenging tasks. They are also empathetic, as they reflect students' emotional states. The results suggest that, across two studies, it was the male learning companion that produced the most positive impact on female students' state-based emotions, attitudes and learning. It is possible that girls transfer their stereotypes to the computer software.

We want to explore the relation between affective states, brainwaves and the learner answers during a multi-choice test questions. 24 participants were used in our experiment. While we were measuring their brainwaves, we asked them to answer 35 questions related to the 7 texts they read, for the first time, the day before. During the experiment, the participants can rate, at any time, their emotional dimensions (pleasure, arousal and dominance) on the Self-Assessment Manikin scale (SAM). Measuring the brainwaves determines the learner mental state and the emotional dimensions indicate the learner affective state. When a participant answers, he mentions if he knows the answer or not. Each answer can be either Right or False. The hypothesis of this paper is: “We can predict the learner's answers from his emotional dimensions and his brainwaves”. By using some machine learning techniques, we reached 90.49% accuracy. In a future work, these results will be implemented in an agent to improve the pedagogical strategies and the adaptation of the content within an Intelligent Tutoring System (STI).

We explored the complex interplay between students' affective states and problem solving outcomes. We conducted a study where 41 students solved 28 analytical reasoning problems from the Law School Admission Test. Participants viewed videos of their interaction history and judged their emotions at theoretically relevant points in the problem solving session (after new problem is displayed, in the midst of problem solving, after feedback is received). We explore excitatory and inhibitory relationships between the affective states and problem solving outcomes (i.e. success or failure, and associated positive or negative feedback). We isolate affective states that are consequences of outcomes and associated feedback as well as affective states that are antecedents of positive or negative outcomes. Follow-up analyses focused on cyclical patterns that incorporate complex relationships between the affective states and problem solving outcomes. Implications of our results for affect-sensitive artificial learning environments are discussed.

In recent years, there has been an increase in research on privacy preserving e-Learning. As authors argue the need for privacy, there were no studies performed to corroborate this need. However, it is established that the learner's emotion do affect his learning. Therefore, this paper investigates the impact of privacy on the learner's emotional state. Particularly, we perform an experiment where participants perform web-based tests, in a no privacy and privacy enforced environments. We report as well on our analysis and findings.

Researchers of educational technologies are often asked to do the impossible: make students learn and have them enjoy it. These two objectives, though not mutually exclusive, are frequently at odds with each other. Effective learning strategies require active knowledge use on the part of the student. Meanwhile, students typically seek to learn through the path of least effort. This can cause conflict during system interaction, and it is often the case that attitudes toward the learning environment suffer. The current study indicates that students' prior expectations of what technology can (or cannot) do may actually have a greater impact than their initial level of motivation, previous domain knowledge, and familiarity with technology, combined. Knowing these prior expectations may be a crucial step to help researchers perform the impossible.

This paper reviews and integrates research that would be necessary to develop an AIED system able to detect and then appropriately react to an affective state of a learner. It addresses the nature of affect, methods to automatically detect affect, as well as the interplay between affect and learning-related cognition, and affective strategies that promote quality learning.

The authors have developed a web-based production system that users can use whenever and anywhere by the Internet. The authors held two cognitive science introductory classes with the system. In our class activities, participants were required to construct running cognitive models on the production system architecture that can solve pulley problems. The participants not only constructed cognitive models but also produced original problems, which were distributed to the other class members. The participants found the defects in their models while trying to solve the problems from other members and trying to improve their models. A posttest indicated that the participants who successfully constructed high performance models revealed deeper understanding of pulley systems.

Recent years have seen increasing interest in narrative-centered learning environments. However, the same qualities that make them engaging can also introduce seductive details that invite off-task behavior. This paper examines off-task behavior in the CRYSTAL ISLAND narrative-centered learning environment. Results from an empirical study examining the relationships between student test performance, individual differences, and off-task behavior are presented. The study found negative correlations between off-task behavior and test performance, as well as significant gender effects on the total amount of off-task behavior. Initial conclusions from a path analysis conducted on students' action sequences are also presented.

A description of a novel domain-independent framework that automatically generates and fades scaffolding supports for task-oriented learning within exploratory environments. Both student knowledge and tutorial assistance is modeled through plan-based interactive narrative.

This paper describes our initial efforts at implementing a new Choice-Adaptive Intelligent Learning Environment (CAILE) that combines multi-agent adaptive technologies and service architectures to provide a framework for designing extendible and reconfigurable learning environments. We describe the core components of the CAILE architecture, learning tasks that establish a situated context for learning, and a set of customizable agents that support student learning. We employ software engineering metrics to evaluate the system, and illustrate the reconfigurable and extensible properties of our design and implementation.

We explore the frequency and impact of misunderstandings in an existing corpus of tutorial dialogues in which a student appears to get an interpretation that is not in line with what the system developers intended. We found that this type of error is frequent, regardless of whether student input is typed or spoken, and that it does not respond well to general misconception repair strategies. Further we found that it is feasible to detect misunderstandings and suggest alternative strategies for repairing them that we intend to test in the future.

We analyze the requirements for an educational Question Answering (QA) system operating on social media content. As a result, we identify a set of advanced natural language processing (NLP) technologies to address the challenges in educational QA. We conducted an inter-annotator agreement study on subjective question classification in the Yahoo!Answers social Q&A site and propose a simple, but effective approach to automatically identify subjective questions. We also developed a two-stage QA architecture for answering learners' questions. In the first step, we aim at re-using human answers to already answered questions by employing question paraphrase identification [1]. In the second step, we apply information retrieval techniques to perform answer retrieval from social media content. We show that elaborate techniques for question preprocessing are crucial.

Identifying effective tutorial strategies is a key problem for tutorial dialogue systems research. Ongoing work in human-human tutorial dialogue continues to reveal the complex phenomena that characterize these interactions, but we have not yet seen the emergence of an automated approach to discovering tutorial dialogue strategies. This paper presents a first step toward establishing a methodology for such an approach. In this methodology, a corpus is first annotated with dialogue acts that are grounded in theories of tutoring and natural language dialogue. Hidden Markov modeling is then applied to discover tutorial strategies inherent in the structure of the sequenced dialogue acts. The methodology is illustrated by demonstrating how hidden Markov models can be learned from a corpus of human-human tutoring in the domain of introductory computer science.