A Survey on Computational Humour From Humour Generation Towards Humour Appreciation

PhD Qualifying Examination


Title: "A Survey on Computational Humour From Humour Generation Towards Humour 
Appreciation"

by

Mr. Andrew CATTLE


Abstract:

As natural language interfaces become more prevalent, the ability for computers 
to both understand and create humour becomes more important. Humour is a 
ubiquitous part of human communication. It can be used to make one's self more 
likeable or to defuse a tense situation or for just pure entertainment. While 
modern digital virtual assistants such as Alexa, Cortana, Google Now, and Siri 
currently do have the ability to tell jokes, these are typically hard-coded 
Easter eggs which have been written and handpicked by humans.

What makes humour such an exciting challenge is that it requires not only 
linguistic dexterity but also world/domain knowledge. Syntax, phonology, and 
semantics all play a role in making a joke funny. The eld of computational 
humour can be divided into three areas: Humour Generation, Humour Recog- nition 
as a classication task, and Humour Recognition as a ranking task. Early 
research, somewhat counter-intuitively, tended to focus on Humour Generation. 
Although Humour Generation is a di cult task even for humans, researchers have 
typically examined highly formulaic types of humour such as punning rid- dles 
or humorous acronyms. Until recently, research into Humour Recognition had 
typically framed it as a classication task; labelling documents simply as 
humorous or not humorous. While this eld has been fairly successful, identi- 
fying several features which consistently indicate humorous intent, the reality 
is that humour exists on a spectrum and ranges from not funny at all to very 
very funny. This is the motivation behind recent reframing of humour recogni- 
tion as a ranking task, with several works focusing on pairwise relative humour 
judgments between short texts such as New Yorker cartoon captions and Twit- ter 
hashtag wars. This reframing also paves the way for Humour Appreciation, 
although such systems do not yet exist.

In this survey, we summarize the existing research relating to Computational 
Humour. First, we explore the task of Humour Generation and examine the var- 
ious joke contexts employed, such as punning riddles and humorous acronyms, as 
well as their implementations and the quality of their outputs. Second, we 
explore the task of Humour Recognition as a classication task and examine the 
strategies employed to infer humorous intent. Finally, we explore the task of 
Humour Recognition as a ranking task and examine the features and systems which 
make up the current state-of-the-art.


Date:			Wednesday, 26 July 2017

Time:                  	10:00am - 12:00noon

Venue:                  Room 2612A
                         Lifts 31/32

Committee Members:	Dr. Xiaojuan Ma (Supervisor)
 			Prof. Qiang Yang (Supervisor)
 			Dr. Brian Mak (Chairperson)
 			Dr. Yangqiu Song
 			Prof. Bertram Shi (ECE)


**** ALL are Welcome ****