06-20, 14:00–14:40 (Europe/Berlin), Frannz Salon
How dense vector functionality was used to provide several ‘Google-like’ capabilities such as Extractive Answers and knowledge graph search over a large dataset at the EU Publications Office.
In this session you will discover how dense vector functionality was used to enhance traditional search and provide users with a 'Google-like' search experience during a proof of concept over a large dataset of multi-lingual legal content curated by the European Union Publications Office in Luxembourg. The presentation will explain how a combination of Elasticsearch, Google BERT transformer models, and the Pureinsights Discovery Platform (PDP) were utilised during the project and discuss the results obtained. There will also be a live demonstration showing the power of semantic understanding across documents and search queries.
Martin is currently Director of International Marketing at Pureinsights and is an evangelist for search and data analytics. He joined Pureinsights from Accenture where he worked for the Search & Content Analytics Group in Applied Intelligence. Prior to Accenture, he was Senior Manager Global OEM Partner Marketing at Qlik. Martin also spent close to 10 years working for the enterprise search vendor Convera.
Martin holds an MBA from Nottingham Business School and a BSc in Mechanical Engineering from Nottingham Trent University.