Multi-Engine Search and Language Translation

Two of the most important elements in user interaction with a database are search and language translation. Search is used to access a database system through queries, for which the accuracy and completeness of response are key challenges. Language translation re-purposes content for a different audience, and the accuracy of translated text can be directly evaluated using search output similarity. In this paper, we summarize previously unpublished approaches to improving the quality of both search and translation, with an aim of improving the accuracy of both of these tasks. Specifically, multi-engine and related metaalgorithmic approaches are shown to be promising means of improving the performance of both search and translation. We then describe the vision of how search and translation can be combined to create a more robust overall text mining project.