Curs 11 Ateptri i satisfacerea lor n parsarea

Curs 11: Aşteptări şi satisfacerea lor în parsarea discursului Dan Cristea Selecţie de sliduri

Expectations and Incremental Discourse Parsing Cristea and Webber, EACL-1998 Cristea et al, Venice-2003 Cristea

Incremental discourse parsing The principle of sequenciality • A left to right reading of

Incremental discourse parsing a TAG inspired approach Adjoining to the right frontier (Polanyi, 1988)

Substitution in case of free expectations k. Although Bill would have wanted it, k+1.

Expectations-driven incremental parsing (Cristea and Webber, 1998) EVIDENCE a. Clinton is bound to win

Expectations-driven incremental parsing a. Clinton is bound to win the elections. b. He is

What can cue-phrases tell us about structure? (Marcu, 1997, 2000; Cristea et al. ,

What can cue-phrases tell us about structure? [Because John is such a generous man

The incremental generation of the first interpretation [Because John is such a generous man

The incremental generation of the second interpretation [Because John is such a generous man

How can references help in discovering the structure? [Because John is such a generous

How can references help in discovering the structure? a. Because Mary was upset, b.

Incremental parallel processing (Cristea et al. , 2005) NP-chunker AR-engine disc. parser segmentator edts-builder

VT guides an incremental discourse parsing (Cristea, 2000; Cristea et al. , 2003, 2005)

The discourse parser implements a „beam search“ N trees observing markers‘ wellformedness constraints reference

Slides: 33

Download presentation

Curs 11: Aşteptări şi satisfacerea lor în parsarea discursului Dan Cristea Selecţie de sliduri

Expectations and Incremental Discourse Parsing Cristea and Webber, EACL-1998 Cristea et al, Venice-2003 Cristea et al, CICLING-2005

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. 6

Incremental discourse parsing a TAG inspired approach Adjoining to the right frontier (Polanyi, 1988) a * 0 ’ a a a a 1 1 0 1 7

Substitution in case of free expectations k. Although Bill would have wanted it, k+1. John sold his bicycle to somebody else. ’ k+1 k k k+1

Expectations-driven incremental parsing (Cristea and Webber, 1998) EVIDENCE a. Clinton is bound to win the elections. a EVIDENCE b ANT-CONS b. He is a natural born campaigner. c ANT-CONS c. If you hold some position on an issue, d e d. then if Clinton wants to get your vote, e. he will assure you with great sincerity that he holds that position too. 8

Expectations-driven incremental parsing a. Clinton is bound to win the elections. b. He is a natural born campaigner. EVIDENCE * b a 9

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, EVIDENCE a * c b 10

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, EVIDENCE a EVIDENCE b c 11

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, EVIDENCE a EVIDENCE b ANT-CONS c ? 12

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, EVIDENCE a b * ANT-CONS c ? 13

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. he will assure you with great sincerity that he holds that position too. EVIDENCE a EVIDENCE b ANT-CONS c ? 14

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, EVIDENCE a EVIDENCE b ANT-CONS c ? ANT-CONS d ? 16

What can cue-phrases tell us about structure? (Marcu, 1997, 2000; Cristea et al. , 2003, 2005) [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) because <something>, <something> 1 2 1 3 2 1 1 2 2 3 2 1 3 4 3 3 because -, - 4 4 1 2 3 4

What can cue-phrases tell us 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] There are only two trees that can be obtained after considering all constraints: because -, - for example 1 because -, - 4 whenever -, - 2 whenever -, - 1 3 because < 1 2 3 >, < 4 > whenever < 2 >, < 3 > < 1 >, < 2 3 > for example -, - for example 2 4 3 because < 1 >, < 2 3 4 > whenever < 2 3 >, < 4 > < 2 >, < 3 > for example

The incremental generation of the first interpretation [Because John is such a generous man 1] [– whenever he is asked for money, 2] ? because -, - 1 * whenever -, - 2

The incremental generation of the first interpretation [Because John is such a generous man 1] [– whenever he is asked for money, 2] [he will give whatever he has, for example 3] because -, - ? 1 whenever -, - 2 3

The incremental generation of the first interpretation [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] because -, - ? 1 4 whenever -, - 2 3

The incremental generation of the second interpretation [Because John is such a generous man 1] [– whenever he is asked for money, 2] because -, - 1 whenever -, - 2

The incremental generation of the second interpretation [Because John is such a generous man 1] [– whenever he is asked for money, 2] [he will give whatever he has, for example 3] because -, -, - for example 1 whenever -, - 2 * 3

The incremental generation of the second interpretation [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] because -, - whenever -, - 1 -, - for example 2 3 4

How can references help in discovering the 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] because -, - for example 1 whenever -, - 1 4 whenever -, - 2 DEA=1 2 right DEA=1 4 -, - for example 2 3 DEA=1 2 3 wrong 4 3 DEA=1 2 DEA=2 4 DEA=2 3

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 c a b wrong V=(a)c DEA=ac b right c V=(a)(b)c DEA=abc

Incremental parallel processing (Cristea et al. , 2005) NP-chunker AR-engine disc. parser segmentator edts-builder summarizer

VT guides an incremental discourse parsing (Cristea, 2000; Cristea et al. , 2003, 2005) 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

The discourse parser implements a „beam search“ N trees observing markers‘ wellformedness constraints reference score cohesion transitions score coherence N

Related presentations