AI-based search and navigation for book collections

One of the issues with search within specialized text collection is the lack of signals from users (or clickstream), which is a contrast to large internet search systems. For example billions of Google users produce enormous amount of search queries and clicks on results each day, enabling the Google developers to optimize the results to superhuman level. To alleviate this, we demonstrate how AI and NLP (natural language processing) provide information on possible contexts that search query can be related to and thus enables to clarify the search result immediately. In addition, the use of our proprietary ‘semantic map’ technology enables the reader to check what other search terms could be used.

Other advanced NLP approaches, like search for semantically similar documents or automatic splitting of text into distinct topics will be also briefly discussed.

We demonstrate the possibilities on a fully functional demo.  

This lecture will be in English.