Abstract: Knowledge graphs are an increasingly important part of modern enterprises as witnessed by the adoption of knowledge graph development by many companies across the world, from Google's Knowledge Graph to Elsevier's Healthcare Knowledge Graph. Enterprise knowledge graphs are a key tool in AI-driven applications by defining enterprise-specific data semantics in the context of data governance, machine learning and human-computer interaction (chatbots). However, the systematic construction and maintenance of an enterprise knowledge graph is extremely challenging, which requires significant automatic support. In this context, we present our ongoing work on knowledge graph extraction from unstructured enterprise text data. The talk will first present an overview of the context of and challenges in enterprise knowledge graph extraction, followed by a presentation of our open source framework for knowledge graph extraction Saffron.
Bio: Paul Buitelaar is Senior Lecturer and vice-director of the Data Science Institute (DSI) at NUI Galway where he leads a team in Natural Language Processing. Paul is also co-PI of the Insight Centre for Data Analytics where he lead the Smart Enterprise demonstrator program and co-lead the Multi-modal Data Analysis research program, which brings together interdisciplinary research from across the Insight Centre in close collaboration with industry partners. Paul's main research interests are in the development and use of Natural Language Processing methods for knowledge extraction and semantic-based information access. He has been involved in a large number of national and international funded projects in this area.