Indexing, Searching & Retrieval of Mathematical Contents

George Washington University, Washington DC

Investigators

Abstract

Efforts are underway worldwide to create Web-accessible mathematical and scientific digital libraries. To leverage such resources, users should be able to search not only for text, but also for equations, formulas and other math constructs. This project will develop the theory and algorithms for indexing and searching of mathematical expressions, and implement a Web-accessible math search and retrieval system. Also, the project will define and implement a math query language that is intuitive and expressive. This will allow users to enter full or fragmentary mathematical expressions as queries, and retrieve all documents containing the symbolic entities that match the query, even if those entities are represented graphically (e.g., GIF images). The methodology is to build a math search engine on top of a text search engine, and to translate math contents and queries into textual queries by means of novel textualization and structure-linearization-&-normalization techniques. A more powerful but demanding alternative strategy, structural indexing, will be developed in a follow-up project. This unprecedented search capability will provide educators and students, from junior high schools to graduate schools, with an effective knowledge-discovery tool to enhance learning and problem-solving skills in mathematics and mathematical sciences. It will also be a valuable aid to researchers and practitioners in those fields.