This research aims to support collaborative distance learners by demonstrating a new way to analyze online knowledge sharing interactions. Our approach applies Hidden Markov Models and Multidimensional Scaling to analyze and assess sequences of coded online student interaction. These analysis techniques were used to train a system to dynamically recognize (1) when students are having trouble learning the new concepts they share with each other, and (2) why they are having trouble. The results of this research may assist an instructor or intelligent coach in understanding and mediating situations in which groups of students collaborate to share their knowledge.

This research aims to support collaborative distance learners by demonstrating how a probabilistic machine learning method can be used to model and analyze online knowledge sharing interactions. The approach applies Hidden Markov Models and Multidimensional Scaling to analyze and assess sequences of coded online student interaction. These analysis techniques were used to train a system to dynamically recognize (1) when students are having trouble learning the new concepts they share with each other, and (2) why they are having trouble. The results of this research may assist an instructor or intelligent coach in understanding and mediating situations in which groups of students collaborate to share their knowledge.

Monitoring and analysis of activities of small groups of students- collocated or at a distance- engaged in synchronous collaborative problem solving activity is the subject of this paper. This is discussed in the frame of Synergo, a new synchronous collaboration support environment that monitors the activity and permits visualization of various quantitative parameters, like density of interaction, symmetry of partners ’ activity, degree of collaboration etc, particularly useful for understanding the mechanics of collaboration. Synergo has been used for synchronous building of flow charts, concept maps, entity-relation diagrams and other semantic modeling activities by small groups of students and has been proposed as a testbed for micro-analysis of small scale interaction in order to gain an insight in collaborative learning.

Abstract. We describe a situation of distance learning based on collaborative production occurring within groups over a significant time span. For such a situation, we suggest giving priority to monitoring and not to guiding systems. We also argue that we need models which are easily computable in order to deal with the heterogeneous and the large scale amount of data related to interactions, i.e. models relying on theoretical assumptions which characterise the structures of groups and of interactions. Social Network Analysis is a good candidate we applied to our experiment in order to compute communication graphs and cohesion factors in groups. This application represents an essential part of a system which would enable tutors to detect a problem or a slowdown of group interaction. 1

Students bring to a collaborative learning situation a great deal of specialized knowledge and experiences that undoubtedly shape the collaboration and learning processes. How effectively this unique knowledge is shared and assimilated by the group affects both the process and the product of the collaboration. In this paper, we describe a machine learning approach, Hidden Markov Modeling, to analyzing and assessing on-line knowledge sharing conversations. We show that this approach can determine the effectiveness of knowledge sharing episodes with 93% accuracy, performing 43% over the baseline. Understanding how members of collaborative learning groups share, assimilate, and build knowledge together may help us identify situations in which facilitation may increase the effectiveness of the group interaction.

A potentially powerful way to aid in the authoring of intelligent tutoring systems is to directly leverage student interaction log data. While problem-solving data has been used in the past to guide the development of tutors, such data has not typically been used as a means to directly construct an initial tutoring system model. We propose an approach called bootstrapping novice data (BND) in which a problem-solving tool is integrated with tutor development software through log files and that integration is then used to create the beginnings of a tutor for the tool. We describe an initial implementation of the BND approach in which Cool Modes, a collaborative software tool, is integrated with the Behavior Recorder, tutor-authoring software that supports development by demonstration. A key to this implementation is a component-based approach in which complementary pieces of software are integrated with little or no change to either software component. We argue that more tutors could be built, and with substantial time savings, using this approach. We discuss some of the lessons learned from this initial effort and from applying the component-based approach, as well as some data analyses that could eventually be performed using the data collected during BND.

The rapid advance of distance learning and networking technology has enabled universities and corporations to reach out and educate students across time and space barriers. Although this technology enables structured collaborative learning activities, online groups often do not enjoy the same benefits as face-to-face learners, and their instructors often do not have time to actively support and mediate the online collaboration. This article demonstrates our capacity to computationally model, analyze, and support online student interaction, in particular knowledge sharing. A unique combination of qualitative analysis and artificial intelligence methods was designed to (a) recognize when students are having trouble learning the new concepts they share with each other, and (b) understand why they are having trouble, so that we might assist an instructor or intelligent coach in mediating group knowledge sharing activities.

In collaborative problem solving, metacognition not only covers strategic reasoning related to the task but also reasoning related to the interaction itself. The hypothesis underlying this work states that regulation of the interaction and regulation of the task are closely related mechanisms and that their co-occurrence facilitates collaborative problem solving. These assumptions are tested experimentally with a traffic simulator. The results show that co-occurrence of task and interaction regulation allow quicker solving of the problem, thus better performance. The experimental treatment aims at observing the effects of interaction meters on the accuracy of subjects' estimation of their participation. Interaction meters are visualization tools that represent the number of contributions related to the discussion and to the implementation of the solution.

In this paper we discuss how group processes can be influenced by designing specific tools in computer supported collaborative leaning. We present the design of a shared workspace application for co-constructive tasks that is enriched by certain functions that are able to track, analyze and feed back parameters of collaboration to group members. Thereby our interdisciplinary approach is mainly based on an integrative methodology for analyzing collaboration behavior and patterns in an implicit manner combined with explicit surveyed data of group members' attitudes and its immediate feedback to the groups. In an exploratory study we examined the influence of this feedback function. Although we could only analyze ad-hoc groups in this study, we detected some benefits of our methodology which might enrich real life Learning Communities' collaboration processes. The data analysis in our study showed advantages of this feedback on processes of a group's well-being as well as parameters of participation. These results provide a basis for further empirical work on problem solving groups that are supported by means of parallel interaction analysis as well as its re-use as information resource.

Interaction analysis is a core function for the support of coaching and evaluation in CSCL. It relies on information captured from the actions performed by the participants during the collaborative process. This information includes data of distinct nature and format, which demands a flexible and standardised data representation, adaptable to different analytical perspectives and collaborative situations. Besides this, it is known that the correct interpretation of human action needs to take context into account. We propose in this paper our approach towards the definition of an XML-based representation of source data, which includes a description of the context of collaboration, and offers a common representation for data of different origin and nature. It is extensible, and independent of the subsequent data analysis methods to which it might be applied. The paper also discusses the possibilities and limitations of XML as a representation language.