Advances in Math Search: Summarization and Relevance Ranking of Math Hits

As digital libraries of mathematical contents become available, one essential support service is math-aware search. The search system must understand math symbols and structures, and allow users to enter queries that involve not only text keywords but also mathematical expressions. In addition, math search must present the query hits in a form that makes it easy for the user to identify quickly the truly relevant hits. To meet that requirement, the hits have to be sorted according to domain-appropriate relevance criteria, and each hit ought to be presented with a query-relevant summary to its target. The standard methods for doing so in text search turned out to be inadequate in math search, and new methods are called for.

As part of the DLMF project, a math search system has been created at NIST. The system builds on text search technology, and breaks new grounds to achieve math-awareness, relevance ranking, and query-relevant summarization/description of hits. This talk will present the math search system, addressing the issues that had to be resolved and the new methods that had to be developed to make the system effective. The system and its novel features, especially relevance ranking and summarization, will be demonstrated live throughout the talk.

Note: Visitors from outside NIST must contact Robin Bickel; (301) 975-3668; at least 24 hours in advance.