Debate iPhone vs Blackberry iPhone of course Blackberry is now for the senior businessmen market The iPhone incarnate the 21st century whereas Blackberry symbolizes an outdated technology The iPhone can reach a very diversified clientele ID: 532571
Download Presentation The PPT/PDF document "Recognizing Stances in Online Debates" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Recognizing Stances in Online Debates
Debate: iPhone vs. BlackberryiPhone of course. Blackberry is now for the senior businessmen market! The iPhone incarnate the 21st century whereas Blackberry symbolizes an outdated technology. The iPhone can reach a very diversified clientele …
ACL 2009 swapna@cs.pitt.eduSlide2
Recognizing Stances in Online Debates
Debate: iPhone vs. BlackberryiPhone of course. Blackberry is now for the senior businessmen market! The iPhone incarnate the 21st century whereas Blackberry symbolizes an outdated technology. The iPhone can reach a very diversified clientele …
ACL 2009 swapna@cs.pitt.edu
Side Classification: pro-iPhone stanceSlide3
Recognizing Stances in Online Debates
Debate: iPhone vs. Blackberry
iPhone
of course
. Blackberry is now for the senior businessmen market!
The
iPhone
incarnate the 21st century
whereas Blackberry symbolizes an outdated technology.
The iPhone can reach a very diversified clientele …
ACL 2009 swapna@cs.pitt.edu
Arguing why their stance is correctSlide4
Recognizing Stances in Online Debates
Debate: iPhone vs. Blackberry
iPhone
of course.
Blackberry is now for the senior businessmen market!
The
iPhone
incarnate the 21st century whereas
Blackberry symbolizes an outdated technology
. The iPhone can reach a very diversified clientele …
ACL 2009 swapna@cs.pitt.edu
Justifying why the opposite side is not goodSlide5
ACL 2009 swapna@cs.pitt.edu
http://www.convinceme.net/Slide6
ACL 2009 swapna@cs.pitt.edu
http://www.convinceme.net/
Side Classification: pro-iPhone stance
Side Classification: pro-Blackberry stance
Side Classification: pro-iPhone stance
Topics:
iPhone
Blackberry
Sides/ Stances:
Pro-iPhone
Pro-Blackberry
Dual-topic,
Dual-sided debates regarding Named EntitiesSlide7
Goal
Debate stance recognition using opinion analysis
Learn debating preferences from the web
Exploited in an unsupervised approach
Combines the individual pieces of information to classify the overall stance
ACL 2009 swapna@cs.pitt.eduSlide8
Challenges
ACL 2009 swapna@cs.pitt.eduThe iPhone incarnate the 21st century whereas Blackberry symbolizes an outdated technologySlide9
Challenges
ACL 2009 swapna@cs.pitt.eduThe iPhone incarnate the 21st century whereas Blackberry symbolizes an
outdated technology
Positive
and
negative
opinions are employed to argue for a side
Side Classification: pro-iPhone stanceSlide10
Challenges
ACL 2009 swapna@cs.pitt.eduThe iPhone incarnate the 21st century
whereas Blackberry symbolizes an outdated technology
Positive
and
negative
opinions are employed to argue for a side
Opinions towards both
topics
within a post
Side Classification: pro-iPhone stanceSlide11
Challenges
ACL 2009 swapna@cs.pitt.eduThe iPhone incarnate the 21st century
whereas Blackberry symbolizes an outdated technology
Positive
and
negative
opinions are employed to argue for a side
Opinions towards both
topics
within a post
Side Classification: pro-iPhone stance
+
towards iPhone- towards BlackberrySlide12
Challenges
ACL 2009 swapna@cs.pitt.eduThe iPhone incarnate the 21st century
whereas Blackberry symbolizes an outdated technology
Positive
and
negative
opinions are employed to argue for a side
Opinions towards both
topics
within a post
Side Classification: pro-iPhone stance
+
towards iPhone- towards Blackberry
We need to consider not only positive and negative opinions
but also what they are about (targets)Slide13
Challenges
Pro-blackberryThe Pearl does music and video nicely …
First, you still can't beat the
full QWERTY keyboard
for quick, effortless typing.
Pro-iPhone
Well,
Apple has always been a well known company
.
Its MAC OS
is also a unique thing.ACL 2009 swapna@cs.pitt.eduSlide14
Challenges
Pro-blackberryThe Pearl does music and video nicely …
First, you still can't beat the
full QWERTY keyboard
for quick, effortless typing.
Pro-iPhone
Well,
Apple has always been a well known company
.
Its MAC OS
is also a unique thing.Debate topics are evoked in a variety of ways: Opinions explicitly toward the named topics are not enoughType of Blackberry
Feature of Blackberry
Maker of iPhone
Feature of iPhone
ACL 2009 swapna@cs.pitt.eduSlide15
Challenges
Pro-blackberryThe Pearl does music and video nicely …
First, you still can't beat the
full QWERTY keyboard
for quick, effortless typing.
Pro-iPhone
Well,
Apple has always been a well known company
.
Its MAC OS
is also a unique thing.We need to consider not only opinions towards topics,But also opinions towards aspectsACL 2009 swapna@cs.pitt.eduSlide16
Challenges
Pro-blackberryThe Pearl does music and video nicely …
First, you still can't beat the
full QWERTY keyboard
for quick, effortless typing.
Pro-iPhone
Well,
Apple has always been a well known company
.
Its MAC OS
is also a unique thing.We need to consider not only opinions towards topics,But also opinions towards aspects
Unique Aspects
ACL 2009 swapna@cs.pitt.eduSlide17
Challenges
iPhone and Blackberry, bothHave e-mail facilitiesCan be used to take photosOperate on batteries
Etc.Both sides share aspects
ACL 2009 swapna@cs.pitt.eduSlide18
Challenges
… I love the ability to receive emails from my work account…ACL 2009 swapna@cs.pitt.edu
People expressing positive opinions regarding emails (generally) prefer BlackberrySlide19
Challenges
… I love the ability to receive emails from my work account…ACL 2009 swapna@cs.pitt.edu
Certain shared aspects may be perceived to be better in one side
Email on Blackberry
Value for shared aspects depends on personal preferences
Emailing – pro-Blackberry people will argue via Email+
That is, Email+ is often a strategy for arguing for the pro-Blackberry stance.
Or, Browsing+ for
iPhoneSlide20
Challenges
… I love the ability to receive emails from my work account…ACL 2009 swapna@cs.pitt.edu
We need to find what a preference/dislike for an individual target means towards the debate stance as a whole Slide21
Challenges
While the iPhone looks nice and does play a decent amount of music, it can't compare in functionality to the BB.
Concessionary opinions can be misleading
ACL 2009 swapna@cs.pitt.edu
Side Classification: pro-Blackberry stanceSlide22
Challenges
While the iPhone looks nice and does play a decent amount of music, it can't compare in functionality to the BB.
ACL 2009 swapna@cs.pitt.edu
Side Classification: pro-Blackberry stance
We need to detect and handle concessionary opinionsSlide23
Challenges: Summary
For debate stance recognition we need to:Consider not only positive and negative opinions, but also what they are about (targets).Consider not only opinions towards topics, but also opinions towards aspectsFind what a preference/dislike for an individual target means towards the debate stance as a wholeDetect and handle concessionary opinions
ACL 2009 swapna@cs.pitt.eduSlide24
Challenges: Summary
For debate stance recognition we need to:Consider not only positive and negative opinions, but also what they are about (targets).Consider not only opinions towards topics, but also opinions towards aspectsFind what a preference/dislike for an individual target means towards the debate stance as a wholeDetect and handle concessionary opinions
ACL 2009 swapna@cs.pitt.edu
Turney
, 2002; Pang et al, 2002; Dave et al, 2003; Yu and
Hatzivassiloglou
, 2003, Pang and Lee 2005, Wilson et al 2005, Goldberg and Zhu, 2006, Kim and
Hovy
2006 …Slide25
Challenges: Summary
For debate stance recognition we need to:Consider not only positive and negative opinions, but also what they are about (targets).Consider not only opinions towards topics, but also opinions towards aspectsFind what a preference/dislike for an individual target means towards the debate stance as a wholeDetect and handle concessionary opinions
ACL 2009 swapna@cs.pitt.edu
Hu
and Liu, 2004; Popescu and
Etzioni
., 2005; Bloom et al. 2007,
Stoyanov
and
Cardie
2008;
Xu et al., 2008 …Slide26
For debate stance recognition we need to:
Consider not only positive and negative opinions, but also what they are about (targets).Consider not only opinions towards topics, but also opinions towards aspectsFind what a preference/dislike for an individual target means towards the debate stance as a wholeDetect and handle concessionary opinions
Our Approach
ACL 2009 swapna@cs.pitt.eduSlide27
For debate stance recognition we need to:
Consider not only positive and negative opinions, but also what they are about (targets).Consider not only opinions towards topics, but also opinions towards aspectsFind what a preference/dislike for an individual target means towards the debate stance as a wholeDetect and handle concessionary opinions
Our Approach
ACL 2009 swapna@cs.pitt.edu
Adopting from previous work,
Opinion-target pairing using Opinion Lexicons and Syntactic rules Slide28
For debate stance recognition we need to:
Consider not only positive and negative opinions, but also what they are about (targets).Consider not only opinions towards topics, but also opinions towards aspectsFind what a preference/dislike for an individual target means towards the debate stance as a wholeDetect and handle concessionary opinions
Our Approach
ACL 2009 swapna@cs.pitt.edu
Unsupervised system
Learn Associations from web and incorporate these towards stance recognition
Adopting from previous work,
Opinion-target pairing using Opinion Lexicons and Syntactic rules Slide29
For debate stance recognition we need to:
Consider not only positive and negative opinions, but also what they are about (targets).Consider not only opinions towards topics, but also opinions towards aspectsFind what a preference/dislike for an individual target means towards the debate stance as a wholeDetect and handle concessionary opinions
Our Approach
ACL 2009 swapna@cs.pitt.edu
Unsupervised system
Learn Associations from web and incorporate these towards stance recognition
Adopting from previous work,
Opinion-target pairing using Opinion Lexicons and Syntactic rules
Rule-based Concession Handler using PDTB connectivesSlide30
Methodology
Learn associations from web data (weblogs)Process the web data to Find opinion-target pairsAssociate opinion-target pairs with each debate side Utilize the associations to classify debate postsProcess the
debate posts toFind opinion-target pairs in the postHandle concessionary opinions
Optimize over all opinion-targets for a post-level stance classification
ACL 2009 swapna@cs.pitt.eduSlide31
Methodology
Learn associations from web data (weblogs)Process the web data to Find opinion-target pairsAssociate opinion-target pairs with each debate side Utilize the associations to classify debate postsProcess the debate posts to
Find opinion-target pairs in the postHandle concessionary opinionsOptimize over all opinion targets for a post-level stance classification
ACL 2009 swapna@cs.pitt.eduSlide32
Methodology: Learning associations
ACL 2009 swapna@cs.pitt.eduDebate title
Topic1 = iPhoneTopic2 = BBSlide33
Methodology: Learning associations
ACL 2009 swapna@cs.pitt.edu
Web search engine
Debate title
Topic
1
= iPhone
Topic
2
= BB
Yahoo search engine APISlide34
Methodology: Learning associations
ACL 2009 swapna@cs.pitt.edu
Web search engine
Debate title
Topic
1
= iPhone
Topic
2
= BB
Weblogs containing both topics
Yahoo search engine APISlide35
Methodology: Learning associations
ACL 2009 swapna@cs.pitt.edu
Web search engine
Debate title
Topic
1
= iPhone
Topic
2
= BB
Weblogs containing both topicsParser
Stanford parserSlide36
Methodology: Learning associations
ACL 2009 swapna@cs.pitt.edu
Web search engine
Debate title
Topic
1
= iPhone
Topic
2
= BB
Weblogs containing both topicsParser
Parsed web documents
Stanford parserSlide37
Methodology: Learning associations
ACL 2009 swapna@cs.pitt.edu
Web search engine
Debate title
Topic
1
= iPhone
Topic
2
= BB
Weblogs containing both topicsParser
Parsed web documents
Opinion-target pairingSlide38
Methodology: Learning associations
ACL 2009 swapna@cs.pitt.edu
Web search engine
Debate title
Topic
1
= iPhone
Topic
2
= BB
Weblogs containing both topicsParser
Parsed web documents
Opinion-target pairing
Lexicon
like:
+
hate:
-
Wilson et al., 2005Slide39
Methodology: Learning associations
ACL 2009 swapna@cs.pitt.edu
Web search engine
Debate title
Topic
1
= iPhone
Topic
2
= BB
Weblogs containing both topicsParser
Parsed web documents
Opinion-target pairing
Lexicon
Syntactic Rules
like = +
hate = -Slide40
Methodology: Learning associations
ACL 2009 swapna@cs.pitt.edu
Web search engine
Debate title
Topic
1
= iPhone
Topic
2
= BB
Weblogs containing both topicsParser
Parsed web documents
Opinion-target pairing
Lexicon
Syntactic Rules
I like email =
email+
like = +
hate = -Slide41
Methodology: Learning associations
ACL 2009 swapna@cs.pitt.edu
Web search engine
Debate title
Topic
1
= iPhone
Topic
2
= BB
Weblogs containing both topicsParser
Parsed web documents
Opinion-target pairing
Lexicon
Syntactic Rules
I like email =
email+
Associations with topic-polarity
like = +
hate = -Slide42
Topic1+
Topic1-
Topic2-
Topic2+
target
j
+
what does a positive opinion towards a target mean with respect to positive or negative opinions regarding either of the topics
Associations with topic-polarity
ACL 2009 swapna@cs.pitt.eduSlide43
Associations with topic-polarity
For each opinion-topic pair (topic1+, topic1-, topic2+, and topic2-) found in the web documentFind other opinion target pairs (target
jp) in its vicinityFor each opinion-target (targetj
p
) calculate its association with each of the opinion-topics
P(topic
1
+|targetj+) P(topic1-|target
j
+)
P(topic2+|targetj+) P(topic2-|targetj+) ACL 2009 swapna@cs.pitt.edu
P(iPhone+ |email+)
P(iPhone- |email+)P(BB+ |email+)
P(BB- |email+)Slide44
Methodology: Learning associations
ACL 2009 swapna@cs.pitt.edu
Web search engine
Debate title
Topic
1
= iPhone
Topic
2
= BB
Weblogs containing both topicsParser
Parsed web documents
Opinion-target pairing
Lexicon
Syntactic Rules
I like email =
email+
Associations with topic-polarity
P(iPhone- |email+)
P(BB- |email+)
P(iPhone+ |email+)
P(BB+ |email+)
like = +
hate = -Slide45
Blackberry+
Blackberry-
iPhone
-
iPhone
+
Storm-
0.062
0.843
0.06
0.03
Associations with topic-polarity
ACL 2009 swapna@cs.pitt.eduSlide46
Methodology
Learn associations from web data (weblogs)Process the web data to Find opinion-target pairsAssociate opinion-target pairs with each debate side Utilize the associations to classify debate postsProcess the debate posts to
Find opinion-target pairs in the postHandle concessionary opinionsOptimize over all opinion targets for a post-level stance classification
ACL 2009 swapna@cs.pitt.eduSlide47
Methodology: Stance Classification
ACL 2009 swapna@cs.pitt.edu
Debate Post
Debate Post
Debate PostSlide48
Methodology: Stance Classification
ACL 2009 swapna@cs.pitt.edu
Debate Post
Debate Post
Debate Post
Parser
Debate Post
Debate Post
Parsed Debate PostSlide49
Methodology: Stance Classification
ACL 2009 swapna@cs.pitt.edu
Debate Post
Debate Post
Debate Post
Parser
Debate Post
Debate Post
Parsed Debate Post
Opinion-target pairing in the post
Lexicon
Syntactic Rules
I like email =
email+Slide50
Methodology: Stance Classification
ACL 2009 swapna@cs.pitt.edu
Debate Post
Debate Post
Debate Post
Parser
Debate Post
Debate Post
Parsed Debate Post
Opinion-target pairing in the post
Lexicon
Syntactic Rules
I like email =
email+
Association lookup, Side Mapping Slide51
Methodology: Stance Classification
ACL 2009 swapna@cs.pitt.edu
Debate Post
Debate Post
Debate Post
Parser
Debate Post
Debate Post
Parsed Debate Post
Opinion-target pairing in the post
Lexicon
Syntactic Rules
I like email =
email+
Association lookup, Side Mapping
P(iPhone- |email+)
P(BB- |email+)
P(iPhone+ |email+)
P(BB+ |email+)Slide52
Topic1+
Topic1-
Topic2-
Topic2+
target+
Association of positive opinion towards a target to positive or negative opinions regarding either of the topics
Association Lookup
0.1
0.05
0.5
0.35
ACL 2009 swapna@cs.pitt.eduSlide53
Side-1
Side-2
Topic1+
Topic1-
Topic2-
Topic2+
target+
Side-1 = Topic1+ alternatively Topic2-
Side-2 =Topic2+ alternatively Topic1-
Association Lookup, Side Mapping
0.1
0.05
0.5
0.35
ACL 2009 swapna@cs.pitt.eduSlide54
target+
Side-1
Side-2
0.15
0.85
Association of positive opinion towards a target to both of the stances
Association Lookup, Side Mapping
ACL 2009 swapna@cs.pitt.eduSlide55
Methodology: Stance Classification
ACL 2009 swapna@cs.pitt.edu
Debate Post
Debate Post
Debate Post
Parser
Debate Post
Debate Post
Parsed Debate Post
Opinion-target
pairingin
the post
Lexicon
Syntactic Rules
I like email =
email+
Association lookup, Side Mapping
P(iPhone- |email+)
P(BB- |email+)
P(iPhone+ |email+)
P(BB+ |email+)Slide56
Methodology: Stance Classification
ACL 2009 swapna@cs.pitt.edu
Debate Post
Debate Post
Debate Post
Parser
Debate Post
Debate Post
Parsed Debate Post
Opinion-target
pairingin
the post
Lexicon
Syntactic Rules
I like email =
email+
Association lookup, Side Mapping
P(iPhone- |email+)
P(BB- |email+)
P(iPhone+ |email+)
P(BB+ |email+)
Assoc(Side-1, email+)
Assoc(Side-2, email+)Slide57
Methodology: Stance Classification
ACL 2009 swapna@cs.pitt.edu
Debate Post
Debate Post
Debate Post
Parser
Debate Post
Debate Post
Parsed Debate Post
Opinion-target pairing in the post
Lexicon
Syntactic Rules
I like email =
email+
Association lookup, Side Mapping
P(iPhone- |email+)
P(BB- |email+)
P(iPhone+ |email+)
P(BB+ |email+)
Concession Handling
Assoc(Side-1, email+)
Assoc(Side-2, email+)Slide58
Concession Handling
Detecting concessionary opinions
Find Concession indicators
Discourse connectives from Penn Discourse Treebank (Prasad et al., 2007)
Use simple rules to find the conceded part of the sentence
While
the
iPhone
looks nice
and
does play a decent amount of music, it can't compare in functionality to the BB.
I
like my
music
, and
phone
,
but
I
don't want
to
carry a brick around
in my pocket when I only need my phone.
ACL 2009 swapna@cs.pitt.eduSlide59
Concession Handling
I like my music, and phone
, but I don't want to carry a brick around
in my pocket when I only need my phone.
Conceded opinions
music+
phone+
ACL 2009 swapna@cs.pitt.eduSlide60
Side-2
Pro-Iphone
Side-1
Pro-Blackberry
music+
phone+
1.0
0.509
0.45
Original associations learnt from the web
Concession Handling
ACL 2009 swapna@cs.pitt.eduSlide61
Side-2
Pro-Iphone
Side-1
Pro-Blackberry
music+
phone+
1.0
0.509
0.45
Concession Handling
Associations after concession handling
Conceded opinions are counted for the opposite side
ACL 2009 swapna@cs.pitt.eduSlide62
Methodology: Stance Classification
ACL 2009 swapna@cs.pitt.edu
Debate Post
Debate Post
Debate Post
Parser
Debate Post
Debate Post
Parsed Debate Post
Opinion-target pairing in the post
Lexicon
Syntactic Rules
I like email =
email+
Association lookup, Side Mapping
P(iPhone- |email+)
P(BB- |email+)
P(iPhone+ |email+)
P(BB+ |email+)
Concession Handling
Post-level association
aggregation
Assoc(Side-1, email+)
Assoc(Side-2, email+)Slide63
Side-2
Pro-Iphone
Side-1
Pro-Blackberry
Aggregation
target
1
+
target
2
+
target
3
+
target
4
+
Each opinion-target pair in the post has a bias toward one or the side
0.9
0.7
0.4
0.5
0.1
0.3
0.6
0.5
ACL 2009 swapna@cs.pitt.eduSlide64
Side-2
Pro-Iphone
Side-1
Pro-Blackberry
Aggregation
target
1
+
target
2
+
target
3
+
target
4
+
Each opinion-target pair in the post has a bias toward one or the other side
Optimize the post classification such that
The side assigned to the post maximizes the association value of the post
ACL 2009 swapna@cs.pitt.eduSlide65
Methodology: Stance Classification
ACL 2009 swapna@cs.pitt.edu
Debate Post
Debate Post
Debate Post
Parser
Debate Post
Debate Post
Parsed Debate Post
Opinion-target pairing in the post
Lexicon
Syntactic Rules
I like email =
email+
Association lookup, Side Mapping
P(iPhone- |email+)
P(BB- |email+)
P(iPhone+ |email+)
P(BB+ |email+)
Concession Handling
Post-level association
aggregation
Side=
pro-Topic1
Assoc(Side-1, email+)
Assoc(Side-2, email+)Slide66
Blackberry+
Blackberry-
iPhone
-
iPhone
+
Storm-
0.062
0.843
0.06
0.03
Associations learnt from web data
ACL 2009 swapna@cs.pitt.eduSlide67
Blackberry+
Blackberry-
iPhone
-
iPhone
+
Storm+
0.227
0.068
0.022
0.613
Associations learnt from web data
Both
OpPMI
, and Op-
Pref
agree with each other;
Both learnt the IS-A relationship
ACL 2009 swapna@cs.pitt.eduSlide68
Blackberry+
Blackberry-
iPhone
-
iPhone
+
Keyboard+
0.718
0.0
0.12
0.09
Associations learnt from web data
ACL 2009 swapna@cs.pitt.eduSlide69
Blackberry+
Blackberry-
iPhone
-
iPhone
+
Keyboard-
0.25
0.25
0.125
0.375
Associations learnt from web data
Negative opinions towards keyboards are not useful for side discrimination
0.5
0.5
ACL 2009 swapna@cs.pitt.eduSlide70
Summing Up
Looked at several tasks ranging from purely lexical to discourse classification
Identify subjective words
Classify their senses as subjective or objective
Recognize, in a text or conversation, whether a word is used with a subjective or objective sense
Sense-aware contextual subjectivity and sentiment analysis
Contextual polarity recognition
Discourse-Level Opinion Interpretation
Many ambiguities are involved in interpreting subjective language!Slide71
Summing Up
Many other ambiguities than these!Sarcasm and IronyYeah, he’s just wonderful.You’re no different from the mob! Oh, there’s a big difference, Mrs. De Marco. The mob is run by murdering, thieving, lying, cheating psychopaths. We work for the President of the United States.
[Married to the Mob]Literal versus non-literal language
He is a pain in the neckSlide72
Pointers
Please see http://www.cs.pitt.edu/~wiebe Publications OpinionFinder
Subjectivity lexicon MPQA manually annotated corpus
Tutorials
Bibliography Slide73
Acknowledgements
Subjectivity Research Group, Pittsburgh
Cem Akkaya, Yaw Gyamfi, Paul Hoffman, Josef Ruppenhofer, Swapna Somasundaran, Theresa Wilson
Cornell:
Claire Cardie, Eric Breck, Yejin Choi, Ves Stoyanov
Utah:
Ellen Riloff, Sidd Patwardhan, Bill Phillips
UNT:
Rada Mihalcea, Carmen Banea
NLP@Pitt:
Wendy Chapman, Rebecca Hwa, Pam Jordan, Diane Litman, …Slide74
Thank you
ACL 2009 swapna@cs.pitt.edu