Commonsense Reasoning from the Angle of Eventualities

PhD Thesis Proposal Defence


Title: "Commonsense Reasoning from the Angle of Eventualities"

by

Mr. Hongming ZHANG


Abstract:

Commonsense reasoning has long been a core artificial intelligence (AI) 
problem. However, there has long been a lack of scalable commonsense 
acquisition methods and principle ways of applying the commonsense in the past. 
An important reason is that we do not have a good enough commonsense 
representation methodology. In this talk, I will first introduce why we should 
represent commonsense with higher-order selectional preference over 
eventualities (i.e., events and states) and how we can construct a large-scale 
eventuality-centric commonsense knowledge graph (KG) ASER at low cost. After 
that, I will show that the collected knowledge is indeed commonsense by 
demonstrating the transferability from ASER to other human-crafted commonsense 
KGs. As ASER cannot cover all the events, I will then introduce how we can 
generalize knowledge about observed events to unseen ones. In the end, I will 
propose the last piece of the thesis work: commonsense knowledge based question 
answering (CKBQA), which is a general commonsense inference learning framework. 
We hope that with CKBQA, we can learn a generalizable commonsense inference 
model to apply the commonsense knowledge in ASER for downstream tasks.


Date:			Wednesday, 31 March 2021

Time:                  	10:30am - 12:30pm

Zoom Meeting:
https://zoom.us/j/2312323804?pwd=RENxMVBuZUgveGhaUDNzMEFLQzEzdz09

Committee Members:	Dr. Yangqiu Song (Supervisor)
  			Prof. Dik-Lun Lee (Chairperson)
 			Prof. Fangzhen Lin
 			Prof. Xiaofang Zhou


**** ALL are Welcome ****