This Small Business Technology Transfer (STTR) Phase I research project involves experiments in conducting learning and assessment processes in parallel, by using agent-based and machine learning techniques to achieve the following objectives: (1) to evolve curricular paths through a knowledge domain automatically, (2) to match human peer-learners for collaborative learning experiences, (3) to embody non-human (agent) peers for tutoring interactions, (4) to predict possible learning outcomes resulting from a range of interventions, and (5) to provide animated visualizations of learners as a means of data reporting. The resulting prototype system will demonstrate state-of-the-art in assessment technology and has the potential to change the face of assessment in schools across the country.
As schools nationwide are confronted with increased testing due to 'No Child Left Behind' policies, the prototype developed by the proposed work could become an indispensable tool for use in any elementary school. If the research demonstrates, as predicted, that a parallel approach to learning and assessment can be effective, then state standardized tests could adapt this methodology. Students will not sacrifice regular in-class learning time for test-taking. In addition, the multi-user, distributed nature of the particular application described within the proposal has far-reaching implications, as students can interact with each other across the country, expanding their circle of learning peers not only beyond the walls of their own classrooms and schools, but beyond the borders of their own states.
