Download presentation
Presentation is loading. Please wait.
Published byMaliyah Cruse Modified over 9 years ago
1
Notes on TAG (LTAG) and Feature Structures September 28 2005 AKJ
2
2 S NP V likes NP e S VP S NP V S* think VP V S does S* NP who Harry Bill substitution adjoining who does Bill think Harry likes LTAG: A Derivation
3
3 Constraints on Substitution and Adjoining S NP V likes VP NP Harry requires a singular NP tree to be substituted at the NP node in
4
4 S NP V likes NP e S VP S NP V S* think VP V S does S* NP who Harry Bill Constraints on Substitution and Adjoining can be adjoined to at the root node of because is anchored on an untensed verb think
5
5 S NP V likes NP e S VP S NP V S* think VP V S does S* NP who Harry Bill The tense associated with the root node of comes from the tense associated with and not from likes Constraints on Substitution and Adjoining: Feature Passing
6
6 Feature Structures Feature Structures: Attribute-Value Structures X1 f: a g: b h: c X2 f: a g: b X1 has more information than X2 X1 is more specific than X2
7
7 Feature Structures Values can be atomic or complex cat: NP agreement: number: sing gender: masc person: third Recursion in FS: For LTAG, no recursion For semantics ?
8
8 Feature Structures Feature Structures: Attribute-Value Structures X1 f: a g: a X2 f: a g: X2 has more information than X1 X2 is more specific than X1 X2 subsumes X1 Co-indexing can be across feature structures also
9
9 Unification of feature structures Given two FS, X1 and X2 X3 = X1 U X2 where X3 is the least FG which subsumes both X1 and X2 X3 is obtained by unifying X1 and X2
10
10 Constraints on Substitution Feature Passing S NP V likes VP NP Harry requires a singular NP tree to be substituted in [num: sing] [num: ] [num: sing] [num: ]
11
11 S NP V likes NP e S VP S NP V S* think VP V S does S* NP who Harry Bill The tense associated with the root node of comes from the tense associated with and not from likes Constraints on Adjoining: Feature Passing [t: ut] [t: ] [t: ut] [t: ] [t: pres] [t: ]
12
12 Top and Bottom Feature Structures We need top (t) and bottom (b) feature structures for each node, especially the internal nodes of a tree. Why? For each node we have a top and bottom view from that node– adjoining can pull apart these two views. When the derivation stops then we unify the top and bottom FS at each node. If one of these unifications fails then the derivation crashes.
13
13 Feature Structures and Unification Adjoining as unification X X* X X X tbtb trbrtrbr tfbftfbf t t r brbr b b f tftf
14
14 Feature Structures and Unification :: X X X t trbrtrbr t t r brbr Substitution as unification
15
15 Obligatory (adjoining) constraints John tried to swim S NP V S* tried VP [ t: -] [ t: ] S PRO TO V to VP swim [t: ] [ t: - ] [ t: +] [t: ] [ t: +] [t: ] If nothing is adjoined to then will crash because the top and bottom features at the root node of will not unify
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.