tailieunhanh - Báo cáo khoa học: "Extractive Summaries for Educational Science Content"

This paper describes an extractive summarizer for educational science content called COGENT. COGENT extends MEAD based on strategies elicited from an empirical study with domain and instructional experts. COGENT implements a hybrid approach integrating both domain independent sentence scoring features and domain-aware features. Initial evaluation results indicate that COGENT outperforms existing summarizers and generates summaries that closely resemble those generated by human experts. | Extractive Summaries for Educational Science Content Sebastian de la Chica Faisal Ahmad James H. Martin Tamara Sumner Institute of Cognitive Science Department of Computer Science University of Colorado at Boulder Abstract This paper describes an extractive summarizer for educational science content called COGENT. COGENT extends MEAD based on strategies elicited from an empirical study with domain and instructional experts. COGENT implements a hybrid approach integrating both domain independent sentence scoring features and domain-aware features. Initial evaluation results indicate that COGENT outperforms existing summarizers and generates summaries that closely resemble those generated by human experts. 1 Introduction Knowledge maps consist of nodes containing rich concept descriptions interconnected using a limited set of relationship types Holley and Dansereau 1984 . Learning research indicates that knowledge maps may be useful for learners to understand the macro-level structure of an information space O Donnell et al. 2002 . Knowledge maps have also emerged as an effective computational infrastructure to support the automated generation of conceptual browsers. Such conceptual browsers appear to allow students to focus on the science content of large educational digital libraries Sumner et al. 2003 such as the Digital Library for Earth System Education . Knowledge maps have also shown promise as domain and student knowledge representations to support personalized learning interactions de la Chica et al. 2008 . In this paper we describe our progress towards the generation of science concept inventories as summaries of digital library collections. Such inventories provide the basis for the construction of knowledge maps useful both as computational knowledge representations and as learning resources for presentation to the student. 2 Related Work Our work is informed by efforts to