NP2022_Trekking into the semantic frontier

Users want library discovery to be precise and intuitive. And solutions on the open web have natural language and contextual semantic search capabilities enabling them to understand users and return lots of information. But many of these tools lack precision. Their results reflect popular searches, but what if a user is looking for something unpopular? What if the user is focused on a very specific aspect of a topic, or researching an unknown relationship between two different topics?

This is where libraries have the advantage. Libraries have rich metadata. Discovery services can look for subject tags. So, discovery users aren’t flooded with irrelevant results that were called up because they share keywords with the query. Instead, the discovery service gets what the user wants.

Users should have the best of both worlds: a semantic search capability that combines natural language with discovery’s subject precision. However, while many institutions and companies are working toward achieving semantic search, there are no true standards and definitions of what semantic search is or should be for academic research.

This panel will discuss the problem’s scope, how the library world is addressing this, and suggest how the research community can come together to craft standards.