Traditional search approaches use keywords to find what you are looking for. But keywords are imperfect things: some concepts are known by multiple words, and some words mean more than one thing.

This is why there's room for improvement with Semantic Search: it uses Artificial Intelligence models that try to capture the meaning of sentences in so called vectors - basically transforming text into numbers.

A class of AI models called Transformer models is used to translate text into vectors, which you can think of as coordinates. These coordinates place a document in a space of 'meaning' - documents with similar content end up close to one another, whereas documents with a different meaning (even if the words are very much the same) end up in a different place. For example:

Being able to explore academic work on the basis of content is in itself interesting, but GlobalCampus adds on this by allowing multiple levels of aggregation. What if we query for similar documents, but then add the scores on the level of an author, a journal or an institution? This allows users to find interesting experts, new venues to publish in or partners to collaborate with.