Knowledge Induction Team @ IBM Research AI - overview

About Us

Our mission is to push the boundary of science and technology to induce knowledge graphs automatically from text. Our goal is to drive the research community while at the same time innovating IBM products, with focus on enterprise search and conversational systems. We are part of the IBM Research AI organization and we have offices in multiple locations, including the IBM Watson Global Headquartes in New York City, the IBM Thomas J. Watson Research Center, Yorktown Heights, the MIT-IBM Watson AI Lab, Cambridge, MA and at the IBM Research - Almaden  lab, CA.
Our work focuses on inducing Knowledge Graphs from corpora. We work closely with the product team to enhace enterprise search and conversational system with knowledge induced directly from customer data. At the same time we publish cutting edge research at top conferences.
Our recent work includes
  • Taxonomy Induction
  • Relation Extraction and Linking
  • Knowledge Base Completion
  • Reading Comprehension
  • Question Answering over Tables
  • Semantic Search


Team Photo


Featured Videos 

Alfio giving a talk on "Extending Knowledge Graphs Using Distantly Supervised Deep Nets" at Knowledge Graph Conference 2019.

A more detailed presentation that was presented at the International Semantic Web Conference 2018 can be found in 

Inducing Implicit Relations from Text using Distantly Supervised Deep Nets (





2020 - Check out our recently accepted ACL 2020 papers
2020 - Check out our recently accepted AAAI 2020 papers
2019 - KI team 2020 won a Research Accomplishments for "Automated Knowledge Induction from IT Support Documents for Creating Virtual Assistants"
2019 - Alfio gave a Keynote Talk at the AAAI-20 Workshop on Knowledge Discovery from Unstructured Data in Financial Services
2019 - Alfio gave a Keynote Talk at the Metadata and Semantics Research conference
2018 - ESWC Winner of Best In Use Paper Award
2017 - Best system at ISWC 2017 Semantic Web Challenge