Grammar Implementation with Lexicalized Tree Adjoining Grammars and - PowerPoint PPT Presentation

Grammar Implementation with Lexicalized Tree Adjoining Grammars and Frame Semantics Grammar implementation with XMG: Frames Laura Kallmeyer, Timm Lichte, Rainer Osswald & Simon Petitjean University of Düsseldorf DGfS Fall School, September 21, 2017 SFB 991

Yesterday: Syntax Implementation of syntactic trees S NP ↓ VP V ⋄ = ⇒ 1 class alphanx0v 2 import VerbProjection [] 3 declare ? Subj 4 { 5 ? Subj = Subject []; 6 ? Subj. ? VP = ? VP 7 } Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 2 2

Today: Frames! Frame theories come with two components: atribute-value descriptions atribute-value constraints How to implement both with XMG? Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 3 3

Atribute-value descriptions (recap.) Vocabulary / Signature Atr atributes ( = dyadic functional relation symbols) Rel (proper) relation symbols Type type symbols ( = monadic predicates) Nname node names (“nominals”) } Nlabel node labels Nvar node variables Primitive atribute-value descriptions (pAVDesc) t | p : t | p � q | [ p 1 , . . . , p n ] : r | p � k ( t ∈ Type, r ∈ Rel, p , q , p i ∈ Atr ∗ , k ∈ Nlabel) Semantics ⎡ ⎤ P ⎢ ⎥ P ∶ t [ P [ t ] ] P ⎢ ⎥ 1 t P [ P , Q ]∶ r ⎢ ⎥ ⎢ ⎥ r ⎣ 2 ⎦ Q r ( 1 , 2 ) Q P ⎡ ⎤ ⎢ ⎥ P ≐ Q ⎢ ⎥ P 1 ⎢ ⎥ ⎢ ⎥ P [ P k [ ] ] P ≜ k ⎣ Q 1 ⎦ Q k Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 4 4

Atribute-value formulas (recap.) Primitive atribute-value formulas (pAVForm) k · p : t | k · p � l · q | � k 1 · p 1 , . . . , k n · p n � : r ( t ∈ Type, r ∈ Rel, p , q , p i ∈ Atr ∗ , k , l , k i ∈ Nlabel) Semantics P P k [ P [ t ] ] 1 ] k ⋅ P ∶ t k [ P t ⟨ k ⋅ P , l ⋅ Q ⟩∶ r k k r 2 ] l [ Q l r ( 1 , 2 ) 1 ] k [ P Q k ⋅ P ≜ l ⋅ Q P k 1 ] l [ Q Q l Formal definitions (fairly standard) Set / universe of “nodes” V Interpretation function I : Atr → [ V ⇀ V ] , Type → ℘ ( V ) , Rel → � n ℘ ( V n ) , Nname → V Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 5 5

Atribute-value formulas in XMG � �� � p p t p : t [ p: t ] t � � p p 1 [ p: ? X1, p � q q: ? X1 ] q 1 q � � p 1 p NOT SUPPORTED YET [ p , q ] : r q 2 r q r ( 1 , 2 ) � � p p k [ ] p � k [ p: ? K [] ] k Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 6 6

Atribute-value formulas in XMG � �� � p k p t ? K [ p: t ]; k · p : t t k � � p ? K [ p: ? X1 ]; k p k · p � l · q k 1 � � ? L [ q: ? X1 ] l q 1 q l � � p NOT SUPPORTED YET � k · p , l · q � : r k k p 1 � � r l q 2 l q r ( 1 , 2 ) Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 7 7

Atribute-value formulas in XMG: Example   causation   1 <frame>{   actor 1 2   ? 0 [ causation,   3 actor: ? 1,  theme  2   4 theme: ? 2,     activity     5 cause: [ activity,   0     cause actor 6  1  actor: ? 1,       7 theme: ? 2 ] ,   theme 2     8   effect: ? 4 [ mover: ? 2, � �   mover 9 2 goal: ? 3 ]   effect  4  10 ]}  goal  3   Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 8 8

Atribute-value constraints (recap.) Constraints (general format) ∀ ϕ , ϕ ∈ AVDesc � V , I , g � � ∀ ϕ iff � V , I , g � , v � ϕ for every v ∈ V Notation: ϕ ⇛ ψ for ∀ ( ϕ → ψ ) Horn constraints: ϕ 1 ∧ . . . ∧ ϕ n ⇛ ψ ( ϕ i ∈ pAVDesc ∪ {⊤} , ψ ∈ pAVDesc ∪ {⊥} ) Examples activity ⇛ event (every activity is an event) causation ∧ activity ⇛ ⊥ (there is nothing which is both a causation and an activity) agent : ⊤ ⇛ agent � actor (every agent is also an actor) activity ⇛ actor : ⊤ (every activity has an actor) ... activity ∧ motion ⇛ actor � mover Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 9 9

Atribute-value constraints in XMG activity ⇛ event activity -> event causation ∧ activity ⇛ ⊥ causation activity -> - agent : ⊤ ⇛ agent � actor agent:+ -> agent = actor activity ⇛ actor : ⊤ activity -> actor:+ activity ∧ motion ⇛ actor � mover activity motion -> actor = mover Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 10 10

Atribute-value constraints in XMG: Examples 1 frame -constraints = { 2 activity -> event, activity -> [ actor:+ ] , 3 motion -> event, motion -> [ mover:+ ] , 4 causation -> event, causation -> [ cause:+,effect:+ ] , 5 locomotion -> activity motion } What is the graphical represention of this (“type hierarchy”)? Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 11 11

Atribute-value constraints in XMG: Examples 1 frame -constraints = { 2 activity -> event, activity -> [ actor:+ ] , 3 motion -> event, motion -> [ mover:+ ] , 4 causation -> event, causation -> [ cause:+,effect:+ ] , 5 locomotion -> activity motion } What is the graphical represention of this (“type hierarchy”)? 1 frame - type -hierarchy = { 2 [ event, [ activity, actor:+, [ locomotion ]] , 3 [ motion, mover:+, [ locomotion ]] , 4 [ causation, cause:+, effect:+ ]]} NOT YET SUPPORTED Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 12 11

Atribute-value constraints (recap.) Graphical presentation of constraints event activity motion causation actor ∶ ⊺ mover ∶ ⊺ cause ∶ ⊺ ∧ effect ∶ ⊺ activity ∧ motion translocation onset-causation extended- actor ≐ mover path ∶ ⊺ cause ∶ punctual-event causation bounded-translocation locomotion goal ∶ ⊺ bounded-locomotion Caveat : Reading convention required ! Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 13 12

Implementation exercise with frames 1 implement the large type hierarchy 2 implement two frame descriptions   causation     actor  1    � �   theme 2 bounded-translocation   e 0     goal  activity  x         actor cause  1          theme  2      3 implement the unfication of these two frames ( e = 0 ) Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 14 13

Case study: dative alternation (recap.) Sketch [→ Kallmeyer/Osswald 2013] (1) a. John sent Mary the book. (double object construction) b. John sent the book to Mary. (prepositional object construction) ⎡ ⎤ ⎢ ⎥ a) S ⎢ ⎥ causation ⎢ ⎥ ⎢ ⎥ cause [ activity actor x ] NP [ i = x ] ⎢ ⎥ VP [ e = e ] ⎢ ⎥ ⎢ ⎥ ⎢ ⎡ ⎤ ⎥ e ⎢ ⎢ ⎥ ⎥ ⎢ ⎥ ⎢ change-of-possession ⎥ V ◇ [ e = e ] NP [ i = z ] NP [ i = y ] ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ effect theme y ⎢ ⎥ ⎢ ⎥ ⎣ ⎣ ⎦ ⎦ recipient z ⎡ ⎤ ⎢ ⎥ b) S ⎢ ⎥ causation ⎢ ⎥ ⎢ ⎥ cause [ activity actor x ] NP [ i = x ] ⎢ ⎥ VP [ e = e ] ⎢ ⎥ ⎢ ⎥ ⎢ ⎡ ⎤ ⎥ e ⎢ ⎥ ⎢ ⎥ PP [ prep = to , i = z , e = e ′ ] ⎢ ⎥ ⎢ bounded-translocation ⎥ VP [ e = e ] ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ effect e ′ mover y ⎢ ⎢ ⎥ ⎥ ⎣ ⎣ ⎦ ⎦ V ◇ [ e = e ] NP [ i = y ] goal z Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 15 14

Implementation exercise with frames Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 16 15

Implementation exercise with frames The trick: sharing of variables across dimensions! Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 17 15

Further applications Morpho-semantic interface: modelling verbal prefixation in Russian (Zinova) modelling derivational morphology in English (Andreou & Petitjean) modelling root-patern morphology in Arabic (Petitjean, Samih & Lichte) Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 18 16

Tomorrow Mon: introduction to grammar engineering and XMG Tue: implementing syntax with XMG Wed: implementing semantics with XMG Thu: parsing implemented grammars with TuLiPA Fri: conclusion Kallmeyer, Lichte, Osswald & Petitjean (HHU Düsseldorf) 19 17

[1] Petitjean, Simon, Younes Samih & Timm Lichte. 2015. Une métagrammaire de l’interface morpho-sémantique dans les verbes en arabe. In Actes de la 22e conférence sur le Traitement Automatique des Langues Naturelles , 473–479. Caen, France. http://www.atala.org/taln_archives/TALN/TALN-2015/taln-2015-court-024 . [2] Zinova, Yulia. 2016. Russian verbal prefixation: a frame semantic analysis . Düsseldorf, Germany: Heinrich-Heine-Universität Düsseldorf Dissertation. https://user.phil-fak.uni-duesseldorf.de/~zinova/Thesis.pdf .