Title: Large scale Knowledge Graphs and applications in Search, Maps, and Shopping

Speaker: Shashidhar Thakur

Abstract

Knowledge Graphs, as a technology to represent structured knowledge, are key to applications that revolve around acquisition and exposition of structured data. We will discuss how these can be specialized for a diverse set of application areas and how they enable grounded understanding of language, modeling of user preferences, which allow us to build high value applications like question answering in Search and Assistant, content recommendation across a variety of use cases, and much more.

Slides

Bio

Shashi received his Bachelor's from the Indian Institute of Technology, Mumbai, where he was recently recognized as a Distinguished Alumnus, and PhD from the University of Texas at Austin, both in Computer Science. He is currently VP, GM, for Consumer Shopping at Google. From 2019-2020 he was VP of Engineering at Google Healthcare division, leading a new product area focused on application of technology and AI to challenging healthcare problems, with the intention of bringing safe and data driven health to everyone on the planet. From 2006-2019, as VP of Search at Google, he led the core search group, driving the strategy of Google's most important product, evolving it towards one that is mobile focused, modern in design, focused on direct answers, and full user journeys to help users get efficient information they need to lead fulfilling lives. His specific contributions include the ubiquitous Google Knowledge graph, which is a technology for representing world knowledge, understanding natural language, and answering user questions on Search and Assistant. And, Google Discover, which is the query-less modality of Search, where information is recommended using AI models of user interest and web content.

Prior to this appointment, he was with Synopsys - a leader in semiconductor chip design automation software for 10 years and Stratify- a start-up formed around developing breakthrough technologies in web crawling and document classification.