CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity!...
Transcript of CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity!...
![Page 1: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/1.jpg)
CS 6120/CS 4120: Natural Language Processing
Instructor: Prof. Lu WangCollege of Computer and Information Science
Northeastern UniversityWebpage: www.ccs.neu.edu/home/luwang
![Page 2: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/2.jpg)
Logistics
• Samples questions posted for final exam. Final exam will be open-book. You can bring your laptop but no Internet access allowed.
• Course project presentation starts next Tuesday. All students arerequired to attend all project presentations. Attendance will be taken.• Presentation days: April 9, 12, 16
• Please fill the course evaluation on TRACE.• To encourage you to do so, 1% bonus point will be given upon completion. • You can inform the TAs and instructor via private post on piazza (for tracking
purpose), by April 21.
![Page 3: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/3.jpg)
Discourse
[Some slides borrowed from Yejin Choi, Jacob Eisenstein, Manfred Pinkal, Stefan Thater, and Michaela Regneri]
![Page 4: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/4.jpg)
Discourse and Coherence
• Linguistic structure beyond the sentence?• What makes...• An argument persuasive?• A story suspenseful?• A joke funny?
![Page 5: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/5.jpg)
Discourse and Coherence
• Linguistic structure beyond the sentence?• What makes...• An argument persuasive?• A story suspenseful?• A joke funny?
• Put another way:• Grammaticality is the property that distinguishes well-structured sentences
from random sequences of words.• Coherence has been proposed to play the same role at the multi-sentence
level. But what are the properties of a coherent text?
![Page 6: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/6.jpg)
![Page 7: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/7.jpg)
![Page 8: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/8.jpg)
![Page 9: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/9.jpg)
![Page 10: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/10.jpg)
![Page 11: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/11.jpg)
![Page 12: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/12.jpg)
Coherence
• John hid Bill’s car keys. He was drunk.• John hid Bill’s car keys. He likes spinach.
![Page 13: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/13.jpg)
Coherence
• John hid Bill’s car keys. He was drunk.• John hid Bill’s car keys. He likes spinach.
• Why one is more coherent than the other?• How do you measure it?
![Page 14: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/14.jpg)
Coherence
![Page 15: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/15.jpg)
Coherence
![Page 16: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/16.jpg)
Discourse
• Discourse is a coherent structured group of textual units (e.g., sentences)• Monologues• Speaker/writer + hearer/reader
• Dialogues• Human-human• Human-computer
![Page 17: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/17.jpg)
Discourse exhibits structure
• Writers use linguistic device to make certain discourse structure• e.g., cue phrases, paragraphs, content flow
• Speakers also use linguistic device to make certain discourse structure• e.g., intonation, gesture, cue phrases
• Readers/Listeners comprehend discourse by recognizing this structure
![Page 18: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/18.jpg)
Discourse Relations
• Discourse relations (Coherence relations) specify the relations between sentences or clauses. Due to these relations, two adjacent sentences can look coherent.
• What is the discourse relation between the following two sentences?
![Page 19: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/19.jpg)
Discourse Relations
• Discourse relations (Coherence relations) specify the relations between sentences or clauses. Due to these relations, two adjacent sentences can look coherent.
• What is the discourse relation between the following two sentences?
![Page 20: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/20.jpg)
More Discourse Relations
![Page 21: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/21.jpg)
Exercise
![Page 22: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/22.jpg)
![Page 23: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/23.jpg)
Rhetorical structure theory (RST)• Nucleus – the central unit, interpretable independently. • Satellite – less central, interpretation depends on N
![Page 24: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/24.jpg)
Rhetorical structure theory (RST)• Nucleus – the central unit, interpretable independently. • Satellite – less central, interpretation depends on N
![Page 25: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/25.jpg)
Rhetorical structure theory (RST)• Nucleus – the central unit, interpretable independently. • Satellite – less central, interpretation depends on N• RST TreeBank (Carlson et al., 2001) defines 78 different RST relations,
grouped into 16 classes.
![Page 26: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/26.jpg)
Examples of RST relations
![Page 27: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/27.jpg)
Examples of RST relations
![Page 28: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/28.jpg)
Examples of RST relations
![Page 29: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/29.jpg)
Discourse Parse Tree for an excerpt from Scientific American (Marcu (2000))
![Page 30: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/30.jpg)
Discourse Parse Tree for an excerpt from Scientific American (Marcu (2000))
![Page 31: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/31.jpg)
Discourse Parsing
• Two related problems: • Discourse Segmentation• Discourse Relation Classification
• Automatic discourse parsing is a very hard problem. (open research problem)
![Page 32: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/32.jpg)
Discourse Segmentation
• Loosely speaking, segmenting a given document into a sequence of subtopics. • The unit of segmentation can be a sentence, or a clause, or even a set
of sentences. (depending on how the result of discourse segmentation will be used.)
![Page 33: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/33.jpg)
Discourse Segmentation• Discourse Marker based Approach• Broadcast News Segmentation: suppose you have a transcript of
broadcast news
![Page 34: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/34.jpg)
Discourse Segmentation• Cohesion based Approach (Halliday & Hasan, 1976)
![Page 35: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/35.jpg)
DotPlot Representation
![Page 36: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/36.jpg)
Discourse Marker (Cue Phrase)
• A cue word/phrase is a word or phrase that functions to signal discourse structure, especially by linking together discourse segments. • e.g., although, but, for example, yet, with, and, well, oh
• Discourse Markers are useful for both • Discourse Segmentation • Discourse Relation Classification
![Page 37: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/37.jpg)
Again, Ambiguity!
• Some discourse markers are ambiguous between “discourse use” V.S. “sentential (non-discourse) use”• With its distant orbit, Mars exhibits frigid weather conditions. • We can see Mars with an ordinary telescope.
![Page 38: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/38.jpg)
Again, Ambiguity!
• Some discourse markers are ambiguous between “discourse use” V.S. “sentential (non-discourse) use”• With its distant orbit, Mars exhibits frigid weather conditions. • We can see Mars with an ordinary telescope.
• Some discourse markers can be used more than one discourse relations• “because” can indicate CAUSE, EVIDENCE• “but” can indicate CONTRAST, ANTITHESIS, CONCESSION
![Page 39: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/39.jpg)
Again, Ambiguity!
• Some discourse markers are ambiguous between “discourse use” V.S. “sentential (non-discourse) use”• With its distant orbit, Mars exhibits frigid weather conditions. • We can see Mars with an ordinary telescope.
• Some discourse markers can be used more than one discourse relations• “because” can indicate CAUSE, EVIDENCE• “but” can indicate CONTRAST, ANTITHESIS, CONCESSION
• Some discourse relations can appear without using any discourse markers.
![Page 40: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/40.jpg)
Annotated corpora
• RST Treebank: 385 English newswire documents• RST Spanish Treebank: several hundred documents, apparently
academic abstracts, http://corpus.iingen.unam.mx/rst/corpus_en.html.• Multilingual RST Treebank: 15 parallel technological abstracts, in
English, Spanish, and Basque• CSTNews Corpus: 50 documents in Brazilian Portuguese• SFU Review Corpus: English and Spanish, 400 review documents each
![Page 41: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/41.jpg)
![Page 42: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/42.jpg)
![Page 43: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/43.jpg)
![Page 44: CS 6120/CS 4120: Natural Language Processingwangluxy/archive/neu_courses/...Again, Ambiguity! •Some discourse markers are ambiguous between “discourse use” V.S. “sentential](https://reader036.fdocuments.us/reader036/viewer/2022071412/61091ea2e815cb2a6f5d2f9f/html5/thumbnails/44.jpg)