Design and implementation for literature search and impact-based summaries

Based on the PARADISE (Platform for Applying, Researching and Developing Intelligent Search Engine), we construct a literature search engine with a collection of papers in computer network area. Besides retrieval, the system also provides comments and impact-based summaries of a paper. The comments are extracted from the citing papers and the summary is obtained from the paper sentences impacting on the citing papers. The functions give users a comprehensive knowledge about the paper they want. We first illustrate how to get candidate comment sentences from the citing papers and generate the citation contexts. Then a KL-divergence-based model is introduced to score the similarity between the sentences in the original paper and the citation contexts. The higher the score is, the more likely academic impact the sentence has. The top rank sentences are combined together as the paper´s summary. The comment sentences are also selected for users to compare the details of related paper work. The experiment exposes the effect of the proposed implementation.