The original basis behind this project is to create a text searcher which can identify discrete topics in a text, be able to account for the intersectionality of ideas, and present them in the format of search results which is easily navigable to a human searcher.
For example, if a researcher in international relations was to be interested in understanding the relationship between globalization and inequality, they could consult the relevant chapters in a book. But say the book is very long and general - it would take too long to read the whole thing. They could consult the index but they would still have to look up each individual instance - moreoever, the index may not account for cases where globalization and inequality are both mentioned. In this example, the text searcher's goal would be to not only point to passages that discuss both globalization and inequality, but also delve deeper - separating subtopics such as how one causes the other, different forms of inequality (social or economic etc...).
In essence, the semantic text searcher's job is to identify ideas and meanings, specializing in overlaps, to help readers extract the most meaningful information for their research in the shortest amount of time possible.