EuroLAN 2019 - Thematic threads in chatseurolan.info.uaic.ro/html/profs/materials/Cristea.pdf ·...

Thematic threads in chats

Dan Cristea

“Al.I.Cuza” University of IasiInstitute of Computer Science of the Romanian Academy

Chapter 1:The discourse and its structure

What is discourse?Longman: 1. a serious speech or piece or writing

on a particular subject: Professor Grant delivered a long discourse on aspects of moral theology. 2. serious conversation between people: You can’t expect meaningful discourse when you two disagree so violently. 3. the language used in particular kinds of speech or writing: scientific discourse.

Text versus discourseSyntactically – a discourse is more than a single

sentence.

From Garcia Marquez

A text is not a discourse!

But it becomes a discourse the very moment it is read or heard by a human... or a machine.

Text versus discourse

Time and discourse

Discourse has a dynamic nature

Time axesreal time

discourse time

story time

2 11000 1030800 920

Attentional state theory (AST)(Barbara Grosz & Candence Sidner, 1987)

Models the linguistic structure of the discourse Gives an account on intentions and how they are combinedExplains the shift of attention during discourse interpretationExplains interruptions and flash-backsPuts in evidence a dynamic domain of referentiality3 components: • a linguistic structure• an intentional structure• an attentional state

Rhetorical structure theory

Basics• text span: un uninterrupted linear interval of text• relation: holds between two or more non-overlapping spans• arguments of relations are of a nuclear type and a satellite type

– a nucleus is more important than a satellite (deletion and substitution tests)

– relations: hypotactic (one nucleus + satellites) and paratactic (all nuclear) • scheme: integrates by a relation two or more text spans (like grammar

rules) • RST analysis are trees• they reflect a judge interpretation (therefore could be subjective)

(William Mann and Sandra Thompson, 1987)

1 2 3 4

H = 1 9 *V = 1 9 *

H = 1V = 1 9 *

H = 9V = 1 9 *

H = 1V = 1 9 *

H = 5V = 1 5 9 *

H = 1V = 1 9 *

H = 3V = 1 3 5 9 *

H = 6 7V = 1 5 6 7 9 *

H = 9V = 1 9 *

H = 9V = 1 (8) 9 *

H = 10V = 1 9 10 *

H = 11V = 1 9 10 11 *H = 3

V = 1 3 5 9DRA = 1 3 H = 9

V = 1 (8) 9DRA = 1 8 9

Trees in RST

relations

nuclear

RST schemes

relation

text span: nucleus

text span: satellite

relation

text span: nucleus

RST relationsSubject matter

(informational)

ElaborationCircumstanceSolutionhoodVolitional CauseVolitional ResultNon-Volitional CauseNon-Volitional ResultPurposeConditionOtherwiseInterpretationEvaluationRestatementSummarySequenceContrast

Presentational (intentional)

MotivationAntithesisBackgroundEnablementEvidenceJustifyConcession

RST analysis1. Farmington Police had to help control traffic recently2. when hundreds of people lined up to be among the first applying for jobs at the yet-to-open Marriot Hotel.3. The hotel’s help-wanted announcement – for 300 openings – was a rare opportunity for many unemployed.4. The people waiting in line carried a message of claims that the jobless could be employed if only they

showed enough moxie.5. Every rule has exceptions,6. but the tragic and too-common tableaux of hundreds of people snake-lining up for any task with a paycheck

illustrates a lack of jobs,7 not laziness.

circumstance

volitional result

evidence

antithesis

concession

background

How do themes relate with trees?

Danes…

Downward: refinement/elaboration of an original idea

Upward: uprising of new aspects out of a previous idea

How do themes relate with trees?

What should a DS count for?

• Unity beyond the sentence level – coherence: it makes sense, with respect to an

accepted setting, real or virtual– why is that certain sequences are more difficult

to read than others?• Referentiality in relation with DS

– cohesion: pronouns and other referential means glues units together

– why is that only certain referential means may be used in certain circumstances?

The Sequentiality Principle (SP)

A left-to-right reading of the terminal frontier of a discourse tree associated to a text corresponds to the span of text analysed, in the same linear order.

(Marcu, 2000), etc.

Operations inspired by TAG

developing treeauxiliary tree transformed tree

foot node

Adjunction

substitution node

developing tree

elementary tree

transformed tree

Substitution

The Right Frontier Constraint• As a referential constraint defining the regions of the

discourse model that could anchor a weak power referential expression

• As an attachment constraint in an incremental development of the tree structure:– at any step of an incremental discourse parsing, if the DT does not

contain substitution nodes, the SP is observed if and only if all operations on the developing DT are adjoinings of left-footed ATsonto nodes of the generalised right frontier or substitutions on the most-inner substitution node.

The Right Frontier Property

Adjunction of a left-footed AT on the RF of the DT

Adjunction of an AT which is not left-footed

σa2σa1

Adjunction of an AT on a node which does not belong to the RF of the DT

σa2σa1

Chapter 2:The structure of chats

Work in collaboration with Corina Forăscu

• Followers of instant messages– In instant messaging each character appeared when it

was typed. The UNIX "talk" command was popular in the 1980s and early 1990s

• Allows easy collaboration, more akin to genuine conversation than email's "letter" format

• Invites to “immediate” reaction from the partner

Why studying chats?• Forums that encourage chats on different

subject matters• Blogs versus chats:

– blogs: short articles displaying your opinion on smth

– chats: direct access with a person (on-line opinion is always more appealing than posting and waiting for reaction)

• Spotting opinions in chats

What makes chats different than direct conversations?

• Direct conversation (face-to-face or telephone) – The emitter has immediate reaction– Although a matter of manners, mainly it is not usual

(offending) to react again before the partner answers– There is no record of previous themes, other than the

memories of partners• Chat

– There is a small delay between messages– The “silence” after sending a short message can be used to

approach another theme– The screen preserves previous themes, which thus can be

re-open easily

RFC can be violated!…

From (Sassen&Kühnlein, 2005)

B: would you say that I’m indeed only jealous? B: do you have another consultation hour scheduled

soon? A: Try to be with yourself after all you’re heading for

your final degree!B: mhm, well, but concentrating on my studies often

doesn’t workA: yes, always on Monday. A: only jealous that would already suffice as problem,

but I think there is more to it.

RFC as conditional

Sassen&Kühnlein 05:

if (not under stress) then RFC

Can RFC be violated?…a. B: I’ve recently felt confused when seeing Maria

talking with Michael. b. B: The reason I’m asking your advise is that I’ll

have my degree in June and right now I’m not in my best shape.

a. A: oh yea, then, what we have here is a nice case of jealousy.

a. B: would you say that I’m indeed only jealous? c. B: do you have another consultation hour scheduled

soon? b. A: Try to be with yourself after all you’re heading

for your final degree!b. B: mhm, well, but concentrating on my studies often

doesn’t workc. A: yes, always on Monday. a. A: only jealous that would already suffice as

problem, but I think there is more to it.

Three different threads develop in parallel

a. B: I’ve recently felt confused when seeing Maria talking with Michael.

b. B: The reason I’m asking your advise is that I’ll have my degree in June and right now I’m not in my best shape.

the jealousy theme

the studies theme

the consultations theme

Could RFC be violated?…a. B: I’ve recently felt confused when seeing Maria

problem, but I think there is more to it.A: to this you have to find a solution immediately.

B: I’ve recently felt confused when seeing Maria talking with Michael.

A: oh yea, then, what we have here is a nice case of jealousy.

B: would you say that I’m indeed only jealous?

A: only jealous that would already suffice as problem, but I think there is more to it.

B: The reason I’m asking your advise is that I’ll have my degree in June and right now I’m not in my best shape.

A: Try to be with yourself after all you’re heading for your final degree!

B: mhm, well, but concentrating on my studies often doesn’t work

B: do you have another consultation hour scheduled soon?

A: yes, always on Monday.

A: to this you have to find a solution immediately.

a1. A: how was your trip to Suceava?b1. A: oh, did I tell you? I have seen yesterday

evening Michael in the disco. b2. A: did you meet him anymore since you two

split?a2. B. Fantastic! lot’s of nice guys and lot of

a: the trip theme

b: the separation from Michael theme

fun. B: eee, but, you know, he was too in that

evening Michael in the disco. b2. A: did you meet him anymore since you two split?a2. B. Fantastic! lot’s of nice guys and lot of fun.

B: eee, but, you know, he was too in that trip!

Merge of two DTsr1

A: how was your trip to Suceava?

B: Fantastic! lot’s of nice guys and lot of fun.

A: did you meet him anymore since you two separated?

A: oh, did I tell you? I have seen yesterday evening Michael in the disco.

B: eee, but, you know, hewas too in that trip!

A: how was your trip to Suceava?

B: Fantastic! lot’s of nice guys and lot of fun.

A: did you meet him anymore since you two separated?

A: oh, did I tell you? I have seen yesterday evening Michael in the disco.

B: eee, but, you know, hewas too in that trip!

Splittingab1. A: so, you didn’t know that I finished with

Michael?!

ab2 A: it happened last month after I came back from Mexico.

a3. B: oh, I’m sorry, are you still mourning after him or you have already someone else?

a4. A: negative! I’m ok but need a period of loneliness.

a5. B: you cannot resist long like this. I know you…

b3. B: so, have you seen the pyramids there?

Splitting a DT

A: so, you didn’t know that I finished with Michael?!

A: it happened last month after I came back from Mexico.

B: oh, I’m sorry, are you still mourning after him or you have already someone else?

A: negative! I’m ok and need a period of loneliness.

B: you cannot resist long like this. I know you…

B: so, have you seen the pyramidsthere?

Splitting a DT

B: so, have you seen the pyramidsthere?

A problem…

How could Mexico be evoked by there since it belongs to a span which is closed ?

B: so, have you seen the pyramids there?

Referential means

• Weak– pronouns (including zero) – demonstratives

• Strong– proper nouns– noun phrases (bridges or coreferential)

Nothing is really forgotten…

Example of a chat

6-7 8-9

13-1415

21-25 26-41

44 45-49

6-7 8-9

13-1415

21-25 26-41

44 45-49

6-7 8-9

13-1415

21-25 26-41

44 45-49

6-7 8-9

13-1415

21-25 26-41

44 45-49

6-7 8-9

13-1415

21-25 26-41

44 45-49

Conclusions• RFC holds!

– not only because a whole class of dedicated scholars have found examples to support it

– because there is a simple theoretical proof for it

• Discourse should be represented by graphs, not trees

• Not only chats, but long discourses too• Graphs can represent multi-threads:

– splitting a thread onto two – merging two distinct threads onto one– interruptions and flashbacks– there are more developing Right Frontiers on a discourse

graph– hopping from a RF to another needs strong referential

Chapter 3:Veins over discourse structure

Work in collaboration with:Nancy Ide, Laurent Romary, Daniel Marcu,

Valentin Tablan, Corina Forascu et al.

Are you comfortable with finding a referent for it in unit 3?

1. With one year before finishing his mandate as president of the company,

2. Mr. W. Ross has begun to bring about its bankruptcy.

3. There were rumors that he has obtained it by fraud.

Are you comfortable with finding a referent for it in unit 3?

2. Mr. W. Ross has begun to bring about its bankruptcy.

*3. There were rumors that he has obtained it by fraud.

How about now?

1. Mr. W. Ross has begun to bring about the bankruptcy of his company.

2. with one year before finishing his mandate as president.

How about now?

2. with one year before finishing his mandate as president.

How many of you think that she in unit 4 refers John’s mother?

1 John told Mary that he loves her.2. He has never been married 3. and lived until his 40s with his mother. 4. She, on the contrary, was married twice.

Discourse structure modifies the distance between anaphor and

antecedent1 John told Mary that he loves her.2. He has never been married3. and lived until his 40s with his mother. 4. She, on the contrary, was married twice.

antithesis

Discourse structure modifies the distance between anaphor and

antecedent1 John told Mary that he loves her.2. He has never been married3. and lived until his 40s with his mother. 4. She, on the contrary, was married twice.

antithesis

elaboration

A satellite can access a far nucleus…

antithesis

elaboration

… but not another satellite

antithesis

elaboration

A nucleus blocks the access between two satellites…

2. Mr. W. Ross has begun to bring about its bankruptcy. 3. There were rumors that he has obtained it by fraud.

circumstance

background

… and a satellite can access its nucleus

2. with one year before finishing his mandate as president.3. There were rumors that he has obtained it by fraud.

background

circumstance

Veins Model basics

Fundamental assumption in VT (the cohesion claim)

• An inter-unit reference is possible only if the two units are in a structural relation one with the other

• The nucleus-satellite distinction, as a component of the discourse structure, gives indications on the range of referents to which an anaphor can be resolved

The definitions of VT: heads• Head expression of a node: the sequence of

the most important units within the corresponding span of text:– the head of a terminal node: its label– the head of a non-terminal node: the concatenation of

the head expressions of the nuclear daughters• the important units are projected up to the level

where the corresponding span is seen as a satellite

to understand a piece of text in the context of the whole discourse one needs the significant units within the span together with other surrounding units

Vein expression of a node: the sequence of units that are required to understand the span of text covered by the node, in the context of the whole discourse

Heads and veins

H=3H=1

H=2 H=3 H=4

1 2 3 4

V=3V=3

V=1 2 3

V=(1 2) 3

V=(1 2) 3 V=3 4

On cohesion

Types of references

evocative references

-evocative resolution processes:- an anaphor may be resolved to a referent that is not linearly the closest, but only hierarchically the closest - based on associations (pattern matching on morpho-semantic features) - fast- give fluency to the text

Types of references

- post-evocative resolution processes:- are inferential processes developed in memory, - computationally and cognitively slow (compel to more inference load),- require more powerful referencing means (like proper nouns), - are less frequent.

post-evocative references

Domain ofevocative accessibility (DEA)

dea(u) = pref(u, vein(u))

Remind! The vein expression of a terminal node(discourse unit): the sequence of units that are required to understand just that unit, in the context of the whole discourse.

(simplified)

Heads and veins

H=3H=1

H=2 H=3 H=4

1 2 3 4

V=3V=3

V=1 2 3

V=(1 2) 3

V=(1 2) 3 V=3 4

From vein expressions...

1 2 3 4

V=1 2 3

V=1 2 3 V=(1 2) 3 V=3 4

... to Domains of Evocative Accessibility

1 2 3 4

V=1 2 3

V=1 2 3 V=3 4

V=1 2 3

The reason why she can refer Mary but not John’s mother

antithesis

elaboration

elaboration 1

V=1 2 4

The reason why we recuperate with difficulty the antecedent of it

2. Mr. W. Ross has begun to bring about its bankruptcy. 3. There were rumors that he has obtained it by fraud.

circumstance

background

… while here the reference is immediate

2. with one year before finishing his mandate as president.3. There were rumors that he has obtained it by fraud.

background

circumstance

V=1 2 3

Experiment 1: evocative vs post-evocative references

4.40%95.60%318176Total

4.50%95.50%11166Romanian

0.90%99.10%11048French

8.30%91.70%9762English

Outside the veins

On the veins

Total no. of refs

No. of units

Source

The 4.4% exceptions

decreasing evoking

power5.00%pronouns

16.00%common nouns

22.70%proper nouns

56.30%pragmatic

VTType of RE

Experiment 2: potential to establish correct co-reference links

• Compare Linear-k and Discourse-VT-kmodels:– For each k, each re, and each model M

(Linear or VT)• p(M-k,re,DEAk) =

• p(M-k,Corpus) = ∑re ∈Corpus p(M-k,re,DEAk)

1, re can be resolved to antecedents in DEAk

0, otherwise.{

Potentials

70,00%

75,00%

80,00%

85,00%

90,00%

95,00%

0 1 2 3 4 5 6 7 8 9E - D E A s i z e

VT-k Linear-k

Experiment 3: the effort required to find antecedents

• Compare Linear-k and Discourse-VT-kmodels:– For each k, each re, and each model M

(Linear or VT)• e(M-k,re,DEAk) =

• e(M-k,Corpus) = ∑re ∈Corpus e(M-k,re,DEAk)

d<k, the distance between re and the closest antecedent in DEAk

k, if no such antecedent exists.{

Effort: an example

Michael D. Casey

Genetic Therapy Inc.

Mr. Casey

Genetic Therapy Inc.

Mr. Casey

the smaller company

Johnson & Johnson

M. James Barett

chairman

its presidentits

Mr. Casey

Mr. Barett

2 3 4 5 6 7 81 9

1. Michael D. Casey, a top Johnson&Johnson manager, moved to Genetic Therapy Inc., a small biotechnology concern here,

2. to become its presidentand chief operating officer.

3. Mr. Casey, 46 years old, was president of J&J's McNeil Pharmaceutical subsidiary,

4. which was merged with another J&J unit, Ortho Pharmaceutical Corp., this year in a cost-cutting move.

5. Mr. Casey succeeds M. James Barrett, 50, as president of Genetic Therapy.

6. Mr. Barrett remains chief executive officer

7. and becomes chairman.8. Mr. Casey said 9. he made the move to the

smaller company.

Efforts

1 3 5 7 9 11 13 15 17 19 25 35 45 55 65 75 85 95

E - D E A s i z eVT Process Lin Process

The account of VT on coherence

• Veins give a natural way to generalize Centering from local to global

Centering Rule 2: transitions

Cb(u) = Cb(u-1) Cb(u) ≠ Cb(u-1)

Cb(u) = Cp(u)

Cb(u) ≠ Cp(u)ABRUPT

SHIFTRETAINING

SMOOTH SHIFT

CONTINUING

CON > RET > SSH > ASH

1 2 3 4

5 V=1 3 5

V=1 3 5

V=1 2 3 5 V=1 3 5V=1 3 4 5

Vein expressions give „lines of argumentation“

1. John sold his bicycle

3. He obtained a good price for it,

5. Therefore he decided to use the money to go on a trip.

1. John sold his bicycle2. although Bill would have wanted it3. He obtained a good price for it,4. which Bill could not have afforded5. Therefore he decided to use the money to go on a trip.

1 2 3 4

5 V=1 3 5

V=1 3 5

V=1 2 3 5 V=1 3 5V=1 3 4 5

Lines of argumentation

4. which Bill could not have afforded.

1. John sold his bicycle

3. He obtained a good price for it,

4. which Bill could not have afforded

5. Therefore he decided to use the money to go on a trip.

Maximal vein expressions: argumentation lines (al)

1 3 51 3 51 3 55

1 3 41 3 41 3 4 54

1 31 3 53

1 21 21 2 3 52

11 3 51

aldea(u)V(u)u

Evaluating the coherence of a discourse

• A smoothness score:– CONTINUING = 4– RETAINING = 3– SMOOTH SHIFT =2– ABRUPT SHIFT = 1– NO Cb = 0

• A global smoothness score: summing up the score of all units

The second conjecture (on coherence)• The global smoothness score of a discourse

when computed following VT is at least as high as the score computed following CT.

• But segments, as considered by Centering, typically are developed along veins.

• When passing segments frontiers, in a linear reading, transitions are usually abrupt.

• Therefore, what we claim here is that long-distance transitions, as computed along veins, are systematically smoother than accidental transitions at segment boundaries.

Transitions and scores on a linear adjacency metric

J = [John], b = [John's bicycle], B = [Bill], p = [price], m = [the money], t = [a trip])

6/4 = 1.5Global

0231Score

No CbSSHRETASHTrans

-pbbJCb

J, m, tp, BJ, p, bB, bJ, bCf

Transitions and scores on a hierarchical adjacency metric

Global

1Score

ASHTrans

B, bJ, bCf

SSHCON

p, BJ, p, bJ, b

11/4=2.75

p, m, tJ, p, bJ, b

Verifying the second conjecture

2.033521.89327173Total

2.341522.1814265Romanian

2.471162.3510947French

1.38841.257659English

Average VT score

per transition

VT scoreAverage CT score

per transition

CT scoreNo. of transitions

Source

Chapter 4:Incremental Discourse Parsing

Work in collaboration with: Bonnie Webber, Ionut Pistol et al.

Incremental discourse parsing

The principle of sequenciality• A left to right reading of the terminal

frontier of the tree associated with a discourse must correspond to the span of text it analyses in the same left-to-right order.

Incremental discourse parsing -a TAG inspired approach

Adjoining to the right frontier

σ0σ0

τ’a

σaσ1

(Polanyi, 1988)

Substitution in case of free expectations

k. Although Bill would have wanted it,

k+1. John sold his bicycle to somebody else.

Expectations-driven incremental parsing

a. Clinton is bound to win the elections.b. He is a natural born campaigner.c. If you hold some position on an issue,d. then if Clinton wants to get your vote,e. he will assure you with great sincerity that he

holds that position too.

EVIDENCE

ANT-CONS

(Cristea and Webber, 1997)

a. Clinton is bound to win the elections.

EVIDENCE

b. He is a natural born campaigner.

a. Clinton is bound to win the elections.b. He is a natural born campaigner.

EVIDENCE

c. If you hold some position on an issue,

EVIDENCE

ANT-CONS

a. Clinton is bound to win the elections.b. He is a natural born campaigner.c. If you hold some position on an issue,

EVIDENCE

ANT-CONS

d. he will assure you with great sincerity that he holds that position too.

a. Clinton is bound to win the elections.b. He is a natural born campaigner.c. If you hold some position on an issue,d. he will assure you with great sincerity that he holds that position too.

EVIDENCE

ANT-CONS

a. Clinton is bound to win the elections.b. He is a natural born campaigner.c. If you hold some position on an issue,d. then if Clinton wants to get your vote,

ANT-CONS

EVIDENCE

ANT-CONS

a. Clinton is bound to win the elections.b. He is a natural born campaigner.c. If you hold some position on an issue,d. then if Clinton wants to get your vote,e. he will assure you with great sincerity that he holds that position too.

ANT-CONS

EVIDENCE

ANT-CONS

a. Clinton is bound to win the elections.b. He is a natural born campaigner.c. If you hold some position on an issue,d. then if Clinton wants to get your vote,e. he will assure you with great sincerity that he holds that position too.

ANT-CONS

EVIDENCE

ANT-CONS

What discourse markers can tell about structure?

[Because John is such a generous man 1] [– whenever he is asked for money, 2] [he will give whatever he has, for example 3] [– he deserves the “Citizen of the year” award. 4]

Reproduced from (Cristea and Webber,1998)

1 2 3 4

because -, -1 2 3 4

1 3 42

1 42 3

because <something>, <something>

(Marcu, 1997, 2000; Cristea et al., 2003, 2005)

whenever -, -2 3 4

whenever <something>, <something>

-, - for example1 2 3

<something>, <something> for example

There are only two trees that can be obtained after considering all constraints:

-, - for example

whenever -, -

because -, -

-, - for example

whenever -, -

because < >, < >1 2 3 4whenever < >, < >2 3< >, < > for example1 2 3

because < >, < >1 2 3 4whenever < >, < >2 3< >, < > for example2 3

The incremental generation of the first interpretation

[Because John is such a generous man 1]

because -, -

[– whenever he is asked for money, 2]

whenever -, -

because -, -

whenever -, -3

[he will give whatever he has, for example 3]

because -, -

whenever -, -

[he will give whatever he has, for example 3][– he deserves the “Citizen of the year” award. 4]

because -, -

whenever -, -

[he will give whatever he has, for example 3][– he deserves the “Citizen of the year” award. 4]

because -, -

whenever -, -

The incremental generation of the second interpretation

because -, -

whenever -, -

[– whenever he is asked for money, 2][he will give whatever he has, for example 3]

-, - for example

because -, -

whenever -, -

-, - for example

[– he deserves the “Citizen of the year” award. 4]

because -, -

whenever -, -

-, - for example 4

[– he deserves the “Citizen of the year” award. 4]

[Because John is such a generous man 1] [– whenever he is asked for money, 2] [he will give whatever he has, for example 3] [– hedeserves the “Citizen of the year” award. 4]

-, - for example

whenever -, -

because -, -

-, - for example

whenever -, -

How can references help in discovering the structure?

DEA=1 4

DEA=1 2 3DEA=1 2

DEA=2 4

DEA=2 3DEA=1 2

wrongright

How can references help in discovering the structure?

a. Because Mary was upset,b. even if John agreed,c. they didn’t speak to one another for several days.

a b V=(a)cDEA=ac

V=(a)(b)cDEA=abcrightwrong

Incremental parallel processing

NP-chunker AR-engine

segmentator edts-builder

disc. parser summarizer

(Cristea et al., 2005)

VT guides an incremental discourse parsing

The tree resulted after parsing is the one which manifests:– the more natural overall references over the

discourse structure – the smoothest overall CT transitions on veins

(Cristea, 2000; Cristea et al., 2003, 2005)

coherence

cohesionreference score

transitions score

The discourse parser implements a „beam search“

Ntrees observing markers‘ well-

formedness constraints

VT referencesCristea,D.; Ide,N.; Romary,L. (1998): Veins Theory. An Approach to Global

Cohesion and Coherence. In Proceedings of Coling/ACL ‘98, MontrealCristea,D., Ide,N., Marcu,D., Tablan, M.-V. (2000): Discourse Structure and

Co-Reference: An Empirical Study, In Proceedings of The 18th International Conference on Computational Linguistics COLING'2000, Luxembourg

Ide,N., Cristea,D. (2000): A Hierarchical Account of Referential Accessibility.In Proceedings of The 38th Annual Meeting of the Association for Computational Linguistics, ACL'2000, Hong Kong

Seretan,V., Cristea,D. (2002): The use of referential constrains in structuringdiscourse. In Proceedings of The Third International Conference onLanguage Resources and Evaluation, LREC-2002, Las Palmas

Cristea, D. (2005): Motivations and Implications of Veins Theory, in B.Sharp(Ed.). Natural Language Understanding and Cognitive Science, Proceedings of the 2nd International Workshop on Natural LanguageUnderstanding and Cognitive Scienc3, NLUCS 2005, in conjunction withICEIS 2005, Miami, U.S.A., May 2005, INSTICC Press

EuroLAN 2019 - Thematic threads in chatseurolan.info.uaic.ro/html/profs/materials/Cristea.pdf ·...

Documents

Transcript of EuroLAN 2019 - Thematic threads in chatseurolan.info.uaic.ro/html/profs/materials/Cristea.pdf ·...

Next Generation Semantic Web Applications - EUROLAN 2013

Silence is talk: Conversational silence in Australian ...

M profs personas and tov workshop v3

SOUND OF SILENCE - Sagsa Spa - SOUND OF SILENCE... · SOUND OF SILENCE SN p20 Silence CABINETS p12 SN Silence p14 Silence SCREEN p1 Silence PARASTA p24 PARASTA p26 WALL p10 SCHEDA

00287-20050301 law profs

Les Profs et Le Staff

Sentiment and Subjectivity Analysis - EuroLAN 2017eurolan.info.uaic.ro/html/profs/materials/Ide.pdf · Sentiment and Subjectivity Analysis An Overview ... through classification,

Networking Tips for Introverts by Marketing Profs

Marketing profs 9_management_practices

M profs 2012_b2_b_content

M profs kit_pinterest_hf197[1]

Marketing Profs 9 Management Practices

Cif partepara profs.

by Profs. Olivier Roche & Frank Shipper

M profs 27_charts_factbook_g0xtw

How event profs-can-be-mobile-smart

00738-amicus final law profs

Verilog Introduction by IIT Kharagpur Profs- Ppts

Silence - How To Handle with Silence

Presentatie revenue profs hsmai 15 november