Richard Alterman

Wikis to support the “collaborative” part of collaborative learning

Prior research has highlighted the value of using wikis to support learning. This paper makes the case that the wiki has several properties that are particularly amenable for constructing applications that support the “collaborative” part of a variety and range of different time/different place student collaborations. In support of the argument, the paper presents the WikiDesignPlatform (WDP). The WDP supplies a suite of awareness, navigation, communication, transcription, and analysis components that provide additional functionality beyond the standard wiki feature set. Two case studies are presented, which have different coordination, communication, and awareness requirements for the “collaborative” part of the students’ collaborative learning activities. The evidence shows that under both conditions, a prefabricated wiki provides a sufficiently rich intersubjective space that adequately supports the students’ collaborative work.

Autonomous Agents that Learn to Better Coordinate

A fundamental difficulty faced by groups of agents that work together is how to efficiently coordinate their efforts. This coordination problem is both ubiquitous and challenging, especially in environments where autonomous agents are motivated by personal goals.Previous AI research on coordination has developed techniques that allow agents to act efficiently from the outset based on common built-in knowledge or to learn to act efficiently when the agents are not autonomous. The research described in this paper builds on those efforts by developing distributed learning techniques that improve coordination among autonomous agents.The techniques presented in this work encompass agents who are heterogeneous, who do not have complete built-in common knowledge, and who cannot coordinate solely by observation. An agent learns from her experiences so that her future behavior more accurately reflects what works (or does not work) in practice. Each agent stores past successes (both planned and unplanned) in their individual casebase. Entries in a casebase are represented as coordinated procedures and are organized around learned expectations about other agents.It is a novel approach for individuals to learn procedures as a means for the group to coordinate more efficiently. Empirical results validate the utility of this approach. Whether or not the agents have initial expertise in solving coordination problems, the distributed learning of the individual agents significantly improves the overall performance of the community, including reducing planning and communication costs.

Participatory adaptation

Expert users of programs that handle complicated data management problems develop methods for coping with data overload, multi-user cooperation, and real-time situations. These expert methods incorporate domain and/or user interface knowledge. If such methods were inherent in a system, then novice users could benefit from the experts experience, the learning curve would be shortened and a more effective system would result. Defining and implementing a complete set of expert methods at design time is a daunting task. Collecting such information from a systems usage, after it has been deployed, should provide a more accurate database of expert methodologies, Current adaptive systems attempt to capture and automate such features during run-time. However, these systems can never evolve very far beyond their original design, since the adaptations occur within the scope of that design. Our method is to offer the experts usage database as input to the designer, reintroducing the designer in the development cycle after a system has been deployed initially, so that a more effective system can be produced in the next generation.

Using shared representations to improve coordination and intent inference

In groupware, users must communicate about their intentions and aintain common knowledge via communication channels that are explicitly designed into the system. Depending upon the task, generic communication tools like chat or a shared whiteboard may not be sufficient to support effective coordination. We have previously reported on a methodology that helps the designer develop task specific communication tools, called coordinating representations, for groupware systems. Coordinating representations lend structure and persistence to coordinating information. We have shown that coordinating representations are readily adopted by a user population, reduce coordination errors, and improve performance in a domain task. As we show in this article, coordinating representations present a unique opportunity to acquire user information in collaborative, user-adapted systems. Because coordinating representations support the exchange of coordinating information, they offer a window onto task and coordination-specific knowledge that is shared by users. Because they add structure to communication, the information that passes through them can be easily exploited by adaptive technology. This approach provides a simple technique for acquiring user knowledge in collaborative, user-adapted systems. We document our application of this approach to an existing groupware system. Several empirical results are provided. First, we show how information that is made available by a coordinating representation can be used to infer user intentions. We also show how this information can be used to mine free text chat for intent information, and show that this information further enhances intent inference. Empirical data shows that an automatic plan generation component, which is driven by information from a coordinating representation, reduces coordination errors and cognitive effort for its users. Finally, our methodology is summarized, and we present a framework for comparing our approach to other strategies for user knowledge acquisition in adaptive systems.

Some computational experiments in summarization

This paper relates the question of narrative summarization to thickness. Roughly, thickness refers to the density of inferred relations between concepts in the narrative as they persist during the reading of the text. We will describe a summarizer, SSS, that takes a piece of text and simplifies it—dethickens it—to produce a description of the events depicted in that text. SSS bases its summary of the story on an event concept coherence (ECC) analysis produced by a program called NEXUS. This program sorts through the events of the story, introducing the implicit ones, grouping together events based largely on their relative positions in an underlying conceptual network. The resulting ECC representation characterizes the complex of relations that exist between the events in the text by a network of instantiated event concepts from which a summary can be produced. In addition to describing some of the inner workings of SSS and the theory of event concept coherence, this paper will develop a quantitative meas...

Participation and common knowledge in a case study of student blogging

The interaction between participation and the emergence of common knowledge is the subject matter of this paper. A case study of a single class provides the focal point of analysis. During the semester the students participated in a blogging activity. As a result of their participation, the students create and distribute knowledge. The online efforts of the students can be described as participation in both a discourse and knowledge community. At one level, blogging is an activity composed of writing, reading, and commenting, and at a second level, the students share their thoughts in their own voices. At a third level, over the course of the semester, the student posts and commentary form a commons of information that can be mined later in the semester for other kinds of learning activities. Knowledge creation, distribution, and accumulation are analyzed in terms of student participation at both the level of individual events and from the perspective of an ongoing community.

Training towards mastery: overcoming the active user paradox

Few users ever attain mastery with an application. Mastery, the state of knowing how to work efficiently with an application requires an understanding of the underlying conceptual model of the system. The active user paradox is one of the main inhibitors of mastery. In this study, we present HotKeyCoach, a training method designed to insert into the flow of the activity learning events that provide contextually relevant training and help the user to circumvent the active user paradox in the pursuit of application mastery.

Reasoning About a Semantic Memory Encoding of the Connectivity of Events

In artificial intelligence, human understanding of text is modeled by programs that construct representations. A critical question concerns determining the form of the representation (hence, the form of the understanding). This article explores the encoding relationship between the semantic memory and the representation/understanding. An important feature of this relationship is that semantic memory and the constructed representation share the same structure, that is, the structure of the “understanding” is a “copy” of some piece of structure taken from semantic memory. A key idea developed is the notion of conceptual roots, the basic framework of the semantic memory-encoded understanding/representation. The research reported in this article examined the twin themes of “copybased” representation/understanding and the conceptual roots by exploring their roles In several reasoning tasks: The conceptual roots can be used to explain succinctly the connection between any two concept-coherent events in an understanding. The semantic memory encoding provides a measure of importance that quantifies the conceptual emphasis of the understanding; given this measure of importance, It is easy to show that each of the important nodes is either a conceptual root or covered by one of the conceptual roots. A combination of techniques can be used to generate a description of the basic event content of the narrative (i.e., a basic summary): This last task reveals some other interesting properties of the conceptual roots.

Rethinking Autonomy

This paper explores the assumption of autonomy. Several arguments are presented against the assumption of runtime autonomy as a principle of design for artificial intelligence systems. The arguments vary from being theoretical, to practical, and to analytic. The latter parts of the paper focus on one strategy for building non-autonomous systems (the practice view). One critical theme is that intelligence is not located in the system alone, it emerges from a history of interactions among user, builder, and designer over a given set of data as mediated by the system. A second critical theme is that artificially intelligent systems are ongoing projects that must be continuously adapted and revised using joint person-machine efforts.

Discourse analysis techniques for modeling group interaction

This paper presents discourse analysis techniques that model the interaction of a small group of users engaged in same-place/different-time interaction. We analyzed data from Vessel World, our experimental testbed, and formulated a modeling technique based on the recurrence of coordination problems and the structure that users create to handle these problems. Subsequent experiments revealed that our original analysis had failed to capture issues with the cognitive load required to maintain common ground. By tracking references users make to both domain and conversational objects, we were able to extract patterns of information access and model the cognitive load incurred to maintain common ground. The improved model of user interaction was successful in explaining systems designed to support interaction.