Institute of Information Science Academia Sinica
Topic: From Knowledge Graph to Event Evolutionary Graph
Speaker: Prof. Ting Liu (School of Computer Science, Harbin Institute of Technology)
Date: 2017-11-30 (Thu) 14:00 – 16:00
Location: Auditorium 122 at CITI Building
Host: Lun-Wei Ku

Abstract:

Events are one of the core concepts of human society, and people's social activities are usually event-driven. The evolutionary principles and patterns that events occur in partial temporal ordering are very valuable commonsense knowledge. The research targets of current knowledge bases, including Knowledge Graph and Semantic Web, are all entities and their relationships, but not events. In order to illustrate the evolutionary principles and development logics between events, we proposed the novel concept of ‘Event Evolutionary Graph’. EEG is a Directed Cyclic Graph, whose nodes are events, and edges stand for the sequential and causal relations between events. It can provide big support for revealing and discovering event evolutionary principles and human behavior patterns. In this report, we will introduce the Chinese knowledge graph ‘Big Cilin’ developed by Harbin Institute of Technology. And we will also introduce the concept, principles, methods and preliminary results of ‘Event Evolutionary Graph’ that we are developing now.