Intelligent Tutoring with Natural Language Support in the Beetle II System

Abstract

We present Beetle II, a tutorial dialogue system designed to accept unrestricted language input and support experimentation with different tutorial planning and dialogue strategies. Our first system evaluation used two different tutoring policies and demonstrated that Beetle II can be successfully used as a platform to study the impact of different approaches to tutoring. In the future, the system can also be used to experiment with a variety of parameters that may affect learning in intelligent tutoring systems.

From Interactive Open Learner Modelling to Intelligent Mentoring: STyLE-OLM and Beyond

“…There have been impressive achievements in the development of knowledge-centric dialogue planners, such as those adopted in BEETLE (Dzikovska et al 2010). ATLAS (Rose et al 2001).…”

Combining Semantic Interpretation and Statistical Classification for Improved Explanation Processing in a Tutorial Dialogue System

“…In recent years, there has been considerable research on tutorial dialogue systems that accept natural language input and engage in dialogue with students to help them improve their answers [1,4,12,13,15,17,20,23]. Such systems are designed to allow students to express their answers in their own words, thus encouraging knowledge construction and harnessing the power of self-explanation [3].…”

“…Many existing tutorial dialogue systems use hand-crafted semantic interpreters to link natural language input with their domain models, in order to produce fine-grained representations of student input [1,2,4,11,20]. Such symbolic NLP systems can support dynamic feedback generation by implementing a library of abstract tutorial strategies, and then, for each new problem or situation, producing a feedback message tailored to the context by choosing a strategy to use and instantiating it from the information gathered from the student answer (see Section 2).…”

“…As our test environment, we use the Beetle II tutorial dialogue system [4], developed to teach concepts in basic electricity and electronics to students without prior knowledge of the domain. The system provides a three-hour self-contained course where students read pre-prepared instructional materials and interact with a circuit simulator.…”

“…To mitigate robustness issues associated with rule-based processing, the system uses a robust interpretation algorithm and a set of error recovery strategies [6]. This approach is successful on the whole in helping students learn, resulting in significant learning gains between pre-and post-tests [4]. However, natural language interpretation failures are correlated with lower learning gains and lower user satisfaction, and there is substantial room for improvement in interpretation quality [8].…”