Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm,...
Transcript of Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm,...
![Page 1: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/1.jpg)
Presenting TWITTIRÒ-UDAn Italian Twitter Treebankin Universal Dependencies
Alessandra Teresa Cignarellaa,b Cristina Boscob and Paolo Rossoa
a. Universitat Politècnica de Valènciab. Università degli Studi di Torino
![Page 2: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/2.jpg)
Motivation
![Page 3: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/3.jpg)
Motivation
1. Sentiment Analysis and Opinion Mining
![Page 4: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/4.jpg)
Motivation
1. Sentiment Analysis and Opinion Mining→ irony, sarcasm, stance, hate speech, misogyny...
![Page 5: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/5.jpg)
Motivation
1. Sentiment Analysis and Opinion Mining→ irony, sarcasm, stance, hate speech, misogyny...
2. Dealing with social media texts
![Page 6: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/6.jpg)
Motivation
1. Sentiment Analysis and Opinion Mining→ irony, sarcasm, stance, hate speech, misogyny...
2. Dealing with social media texts→ hard!!
![Page 7: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/7.jpg)
Motivation
1. Sentiment Analysis and Opinion Mining→ irony, sarcasm, stance, hate speech, misogyny...
2. Dealing with social media texts→ hard!!
3. Syntax
![Page 8: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/8.jpg)
Motivation
1. Sentiment Analysis and Opinion Mining→ irony, sarcasm, stance, hate speech, misogyny...
2. Dealing with social media texts→ hard!!
3. Syntax→ Universal Dependencies are cool!
![Page 9: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/9.jpg)
Research Questions
![Page 10: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/10.jpg)
Research Questions
1. How can we automatically detect irony ?
![Page 11: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/11.jpg)
Research Questions
1. How can we automatically detect irony ?
2. Could syntax information help in the detection of irony?
![Page 12: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/12.jpg)
Research Questions
1. How can we automatically detect irony ?
2. Could syntax information help in the detection of irony?...and maybe help in other detection tasks too?
![Page 13: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/13.jpg)
Research Questions
1. How can we automatically detect irony ?
2. Could syntax information help in the detection of irony?...and maybe help in other detection tasks too?
Our approach:
![Page 14: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/14.jpg)
Research Questions
1. How can we automatically detect irony ?
2. Could syntax information help in the detection of irony?...and maybe help in other detection tasks too?
Our approach:
Let’s build a corpus and find out!
![Page 15: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/15.jpg)
What is TWITTIRÒ-UD ?
![Page 16: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/16.jpg)
What is TWITTIRÒ-UD ?
Treebank
![Page 17: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/17.jpg)
What is TWITTIRÒ-UD ?
TreebankItalian
![Page 18: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/18.jpg)
What is TWITTIRÒ-UD ?
TreebankItalian
![Page 19: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/19.jpg)
What is TWITTIRÒ-UD ?
TreebankItalian
Universal Dependencies
![Page 20: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/20.jpg)
What is TWITTIRÒ-UD ?
TreebankItalian
Universal Dependencies
Irony
Sarcasm
![Page 21: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/21.jpg)
Related Work
![Page 22: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/22.jpg)
Related Work
Social media & Twitter:
![Page 23: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/23.jpg)
Related Work
Social media & Twitter:● Tagging the Twitterverse (Foster et al., 2011)
● The French Social Media Bank (Seddah et al., 2012)
● TWEEBANK (Kong et al., 2014)
● TWEEBANK v2 (Liu et al., 2018)
● Arabic (Albogamy and Ramsay, 2017)
● African-American English (Blodgett et al., 2018)
● Hindi English (Bhat et al., 2018)
![Page 24: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/24.jpg)
Related Work
![Page 25: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/25.jpg)
Related Work
![Page 26: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/26.jpg)
Related Work
Two main references for our work:
![Page 27: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/27.jpg)
Related Work
Two main references for our work:● UD_Italian treebank (Simi et al., 2014)
![Page 28: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/28.jpg)
Related Work
Two main references for our work:● UD_Italian treebank (Simi et al., 2014)
● PoSTWITA-UD (Sanguinetti et al., 2018)
![Page 29: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/29.jpg)
Data
![Page 30: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/30.jpg)
Data
● 1,424 tweets from TWITTIRÒ (Cignarella et al., 2018)
![Page 31: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/31.jpg)
Data
● 1,424 tweets from TWITTIRÒ (Cignarella et al., 2018)
● fine-grained irony annotation (Karoui et al. 2017)
![Page 32: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/32.jpg)
Data
● 1,424 tweets from TWITTIRÒ (Cignarella et al., 2018)
● fine-grained irony annotation (Karoui et al. 2017)
1. EXPLICIT2. IMPLICIT
![Page 33: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/33.jpg)
Data
● 1,424 tweets from TWITTIRÒ (Cignarella et al., 2018)
● fine-grained irony annotation (Karoui et al. 2017)
1. ANALOGY2. EUPHEMISM3. RHETORICAL QUESTION4. OXYMORON or PARADOX5. FALSE ASSERTION6. CONTEXT SHIFT7. HYPERBOLE or EXAGGERATION8. OTHER
1. EXPLICIT2. IMPLICIT
![Page 34: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/34.jpg)
Data
● 1,424 tweets from TWITTIRÒ (Cignarella et al., 2018)
● fine-grained irony annotation (Karoui et al. 2017)
● sarcasm annotation (EVALITA 2018)
1. ANALOGY2. EUPHEMISM3. RHETORICAL QUESTION4. OXYMORON or PARADOX5. FALSE ASSERTION6. CONTEXT SHIFT7. HYPERBOLE or EXAGGERATION8. OTHER
1. EXPLICIT2. IMPLICIT
![Page 35: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/35.jpg)
Annotation
![Page 36: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/36.jpg)
Annotation
# text = Presentato il nuovo iPhone. È già al 36% di batteria.
![Page 37: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/37.jpg)
Annotation
# text = Presentato il nuovo iPhone. È già al 36% di batteria.
# irony = EXPLICIT OXYMORON/PARADOX
![Page 38: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/38.jpg)
Annotation
# text = Presentato il nuovo iPhone. È già al 36% di batteria.
# irony = EXPLICIT OXYMORON/PARADOX
# sarcasm = 1
![Page 39: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/39.jpg)
Annotation
# text = Presentato il nuovo iPhone. È già al 36% di batteria.
# irony = EXPLICIT OXYMORON/PARADOX
# sarcasm = 1
Translation:The new iPhone has been launched. Battery is already at 36%.
![Page 40: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/40.jpg)
Data
![Page 41: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/41.jpg)
Data
With the tool UDPipe:
![Page 42: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/42.jpg)
Data
With the tool UDPipe:● tokenization● lemmatization● PoS-tagging● dependency parsing
![Page 43: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/43.jpg)
Data
With the tool UDPipe:● tokenization● lemmatization● PoS-tagging● dependency parsing
}
![Page 44: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/44.jpg)
Data
With the tool UDPipe:● tokenization● lemmatization● PoS-tagging● dependency parsing
} 1,424 tweets!(17,933 tokens)
![Page 45: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/45.jpg)
Data
With the tool UDPipe:● tokenization● lemmatization● PoS-tagging● dependency parsing
Full release in the UD repository: November 2019
} 1,424 tweets!(17,933 tokens)
![Page 46: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/46.jpg)
Data
![Page 47: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/47.jpg)
Data
![Page 48: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/48.jpg)
Data
![Page 49: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/49.jpg)
Data
1. Fine-grained annotation for irony
![Page 50: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/50.jpg)
Data
1. Fine-grained annotation for irony
![Page 51: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/51.jpg)
Data
1. Fine-grained annotation for irony2. Morpho-syntactic information
![Page 52: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/52.jpg)
Issues Encountered and Lessons Learned
![Page 53: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/53.jpg)
Issues Encountered and Lessons Learned
● Tokenization errors depending on misspelled words
![Page 54: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/54.jpg)
Issues Encountered and Lessons Learned
● Tokenization errors depending on misspelled words
xkè → perché
![Page 55: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/55.jpg)
Issues Encountered and Lessons Learned
● Tokenization errors depending on misspelled words
● Punctuation irregularly used
xkè → perché
![Page 56: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/56.jpg)
Issues Encountered and Lessons Learned
● Tokenization errors depending on misspelled words
● Punctuation irregularly used
● Twitter marks
xkè → perché
![Page 57: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/57.jpg)
Issues Encountered and Lessons Learned
● Tokenization errors depending on misspelled words
● Punctuation irregularly used
● Twitter marks #hashtag
xkè → perché
![Page 58: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/58.jpg)
Issues Encountered and Lessons Learned
● Tokenization errors depending on misspelled words
● Punctuation irregularly used
● Twitter marks #hashtag
@mention
xkè → perché
![Page 59: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/59.jpg)
Issues Encountered and Lessons Learned
● Tokenization errors depending on misspelled words
● Punctuation irregularly used
● Twitter marks
● No sentence splitting
#hashtag
@mention
xkè → perché
![Page 60: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/60.jpg)
Issues Encountered and Lessons Learned
● Tokenization errors depending on misspelled words
● Punctuation irregularly used
● Twitter marks
● No sentence splitting
● Single-root constraint
#hashtag
@mention
xkè → perché
![Page 61: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/61.jpg)
Issues Encountered and Lessons Learned
![Page 62: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/62.jpg)
Issues Encountered and Lessons Learned
![Page 63: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/63.jpg)
Issues Encountered and Lessons Learned
![Page 64: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/64.jpg)
Issues Encountered and Lessons Learned
![Page 65: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/65.jpg)
Issues Encountered and Lessons Learned
![Page 66: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/66.jpg)
Issues Encountered and Lessons Learned
![Page 67: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/67.jpg)
Other Highlights
![Page 68: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/68.jpg)
Other Highlights
● Punctuation is indeed exploited more extensively in the two social media datasets rather than in UD_Italian.
![Page 69: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/69.jpg)
Other Highlights
● Punctuation is indeed exploited more extensively in the two social media datasets rather than in UD_Italian.
● Mentions and hashtags have a similar distribution in the two social media datasets.
![Page 70: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/70.jpg)
Other Highlights
● Punctuation is indeed exploited more extensively in the two social media datasets rather than in UD_Italian.
● Mentions and hashtags have a similar distribution in the two social media datasets.
● The use of passive voices (aux:pass) is low in PoSTWITA-UD and in TWITTIRÒ-UD, indicating a preference for the exploitation of active voices, as it happens in spoken language.
![Page 71: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/71.jpg)
A Parsing Experiment
![Page 72: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/72.jpg)
A Parsing Experiment
We performed an evaluation of UDPipe using the TWITTIRÒ-UD gold corpus as a test set.
![Page 73: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/73.jpg)
A Parsing Experiment
We performed an evaluation of UDPipe using the TWITTIRÒ-UD gold corpus as a test set.
The following settings were exploited:
![Page 74: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/74.jpg)
A Parsing Experiment
We performed an evaluation of UDPipe using the TWITTIRÒ-UD gold corpus as a test set.
The following settings were exploited:
1. training UDPipe using only UD_Italian
![Page 75: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/75.jpg)
A Parsing Experiment
We performed an evaluation of UDPipe using the TWITTIRÒ-UD gold corpus as a test set.
The following settings were exploited:
1. training UDPipe using only UD_Italian
2. training UDPipe using only PoSTWITA-UD
![Page 76: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/76.jpg)
A Parsing Experiment
We performed an evaluation of UDPipe using the TWITTIRÒ-UD gold corpus as a test set.
The following settings were exploited:
1. training UDPipe using only UD_Italian
2. training UDPipe using only PoSTWITA-UD
3. training UDPipe using both resources
![Page 77: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/77.jpg)
A Parsing Experiment
![Page 78: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/78.jpg)
A Parsing Experiment
![Page 79: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/79.jpg)
A Parsing Experiment
![Page 80: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/80.jpg)
A Parsing Experiment
Results in-line with state of the art(PoSTWITA-UD, Sanguinetti et al., 2018)
![Page 81: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/81.jpg)
Conclusions
![Page 82: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/82.jpg)
Conclusions
● We discuss the annotation of this resource which encompasses a fine-grained representation of irony and the UD morpho-syntactic analysis
![Page 83: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/83.jpg)
Conclusions
● We discuss the annotation of this resource which encompasses a fine-grained representation of irony and the UD morpho-syntactic analysis
● Release of the complete resource (1,424 tweets) to be accomplished in November 2019
![Page 84: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/84.jpg)
Conclusions
● We discuss the annotation of this resource which encompasses a fine-grained representation of irony and the UD morpho-syntactic analysis
● Release of the complete resource (1,424 tweets) to be accomplished in November 2019
● It enriches the scenario of available resources for a text genre which is especially hard to parse (social media texts)
![Page 85: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/85.jpg)
Future Work
![Page 86: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/86.jpg)
Future Work
● Investigation of possible relationships between syntax and semantics of the uses of figurative language (irony in particular)
![Page 87: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/87.jpg)
Future Work
● Investigation of possible relationships between syntax and semantics of the uses of figurative language (irony in particular)→ ongoing experiments...
![Page 88: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/88.jpg)
Future Work
● Investigation of possible relationships between syntax and semantics of the uses of figurative language (irony in particular)→ ongoing experiments...
● A resource whose annotation encompasses both UD relations and a fine-grained description of irony may indeed pave the way for the investigation of whether syntactic knowledge might help in SA and other related tasks
![Page 89: Presenting TWITTIRÒ-UD · Motivation 1. Sentiment Analysis and Opinion Mining → irony, sarcasm, stance, hate speech, misogyny... Motivation 1. Sentiment Analysis and Opinion Mining](https://reader030.fdocuments.in/reader030/viewer/2022041008/5eb3a1cddcd02756c7565568/html5/thumbnails/89.jpg)
Future Work
● Investigation of possible relationships between syntax and semantics of the uses of figurative language (irony in particular)→ ongoing experiments...
● A resource whose annotation encompasses both UD relations and a fine-grained description of irony may indeed pave the way for the investigation of whether syntactic knowledge might help in SA and other related tasks→ new NLP features for Sentiment Analysis?