NASSCOM Annual Technology Conference 2021

About the Workshop:
Open Information Extraction (Open IE) attempts to extract (subject, relation, object) tuples from a sentence in an ontology-independent manner, i.e., all arguments and relations are phrases extracted from the original text. These semi-structured representations are useful for constructing knowledge bases without human supervision and improving question answering systems. Open IE has a 15 year old history in the field of NLP with a wide variety of techniques ranging from rule-based systems to machine learning systems, both non-neural, and recently neural. In this tutorial, we describe Open IE’s philosophy, salient techniques, applications, latest trends and challenges that lie ahead.

Agenda:
10:00-10:25 am: Introduction to OpenIE and its Applications
10:25-10:45 am: Traditional approaches to OpenIE
10:45-11:15 am: Primer on Deep Learning for NLP
11:15-11:45 am: Deep Learning approaches to OpenIE
11:45-12:00 pm: Future of OpenIE