LIRICS WP3: Morphosyntactic and syntactic annotations

1
LIRICS WP3 Morpho-syntactic and syntactic
annotations

• Thierry Declerck
• declerck_at_dfki.de
• DFKI-LT - Saarbrücken
• 23rd May 2006

2
Summary of the presentation

• WP3 1st year objectives
• objectives and work done in T3.1, T3.2
• status of deliverables in T3.1, T3.2
• Main Achievements SynAF NWI accepted and
  Working Draft discussed at ISO level
• Next Steps Data Categories for Syntactic
  Annotation
• Synergies with other LIRICS WPs, ISO activities,
  meetings

3
Task 3.1 Survey

• Task 3.1 Evaluation of initiatives for
  morpho-syntactic and syntactic annotation .
• ObjectivesEvaluate past and current initiatives
  dealing with standardisation, guidelines and
  recommendations for morpho-syntactic and
  syntactic annotation. Background for starting ISO
  standardization activity on syntactic annotation
  (model data categories)
• Results Deliverable 3.1 (V1 and V2).

4
Task 3.1 Deliverables
5
Task 3.2 ISO Proposal for a standard for
Syntactic Annotation

• Submit to ISO TC37/SC4 a New Work Item on
  Syntactic Annotation, and go through the ISO
  procedure till the final document.
• Objectives Proposal for morpho-syntactic
  annotation meta-model standard together with a
  wide coverage of data category selection
  formorpho-syntatic annotation
• Results New Work Item on Syntactic Annotation
  (SynAF) succesfully submitted to ISO TC37/SC4.
  Ongoing Discussion on the Working Draft

6
Task 3.2 Deliverables
Interim Report WD of morpho-syntatic annotation
standard for CD ballot.
CD of morpho-syntatic annotation standard for ISO
DIS ballot
7
The SynAF Proposal

• Syntactic Annotation has 2 Functions in NLP
• 1) To represent linguistic constituencies, like
  Noun Phrases (NP), describing a structured
  sequence of morpho-syntactically annotated items,
  where we consider also constituents built from
  non-contiguous elements, and
• 2) To represent dependency relations dependency
  information can exist between morpho-syntactically
  annotated items within a phrase (an adjective is
  the modifier of the head noun within an NP) or
  describe a specific relation between syntactic
  constituents at the clausal and sentential level
  (i.e. an NP being the "subject" of the main verb
  of a clause or sentence). In the first case we
  speak of an internal dependency and in the second
  case we speak of an external dependency. But the
  dependency relation can also be stated including
  empty elements (like the pro-drop property in
  romance languages)

8
The SynAF Proposal (2)

• SynAF is concerned thus with a meta-model that
  covers both dimensions of syntactic constituency
  and dependency, and SynAF is proposing a
  multi-layered annotation framework that allows
  the combined and interrelated annotation of
  language data along both lines of consideration.
  Also the data-categories to be proposed to ISO
  standardization will be about the basic
  annotation concerning both dimensions.

9
The SynAF Proposal (3)

• A main starting point Tiger
• The Tiger annotation framework foresees 2 types
  of annotation for constituency (represented than
  by a node label in the annotation framework) and
  for dependency (represented as an edge label in
  the annotation framework).

10
The SynAF Proposal (4)

• Another starting point ISST
• The approach followed in the ISST (Italian Syntax
  Semantic Treebank) framework, is similar to the
  one proposed in Tiger, in the sense that a
  multi-layered syntactic annotation strategy is
  proposed One level for constituency and one
  level for dependency, with a pointing mechanism
  for referring from the second level to the first
  one.

11
Example for Tiger

• ltbodygt lts id"s5"gt
• ltgraph root"s5_504"gt
• ltterminalsgt
• ltt id"s5_1" word"Die" pos"ART"
  morph"Def.Fem.Nom.Sg"/gt
• ltt id"s5_2" word"Tagung" pos"NN"
  morph"Fem.Nom.Sg."/gt
• ltt id"s5_3" word"hat" pos"VVFIN"
  morph"3.Sg.Pres.Ind"/gt
• ltt id"s5_4" word"mehr" pos"PIAT" morph"--"/gt
  
• ltt id"s5_5" word"Teilnehmer" pos"NN"
  morph"Masc.Akk.Pl."/gt
• ltt id"s5_6" word"als" pos"KOKOM" morph"--"/gt
  
• ltt id"s5_7" word"je" pos"ADV" morph"--"/gt
  
• ltt id"s5_8" word"zuvor" pos"ADV" morph"--"/gt
  
• lt/terminalsgt

• ltnonterminalsgt
• ltnt id"s5_500" cat"NP"gt
• ltedge label"NK" idref"s5_1"/gt
• ltedge label"NK" idref"s5_2"/gt
• lt/ntgt
• ltnt id"s5_501" cat"AVP"gt
• ltedge label"CM" idref"s5_6"/gt
• ltedge label"MO" idref"s5_7"/gt
• ltedge label"HD" idref"s5_8"/gt
• lt/ntgt
• ltnt id"s5_502" cat"AP"gt
• ltedge label"HD" idref"s5_4"/gt
• ltedge label"CC" idref"s5_501"/gt
• lt/ntgt
• ..
• lt/nonterminalsgt

12
Example for ISST

• ltfrase id"0" morfofile"sole.morph026"
  rs"Presentato un libro bianco del Governo Major
  ."gt
• ltnodo tipo"F3"gt
• ltnodo tipo"SV3" id"0"gt
• ltfoglia lemma"presentare"
  href"mw_001"/gt
• ltnodo tipo"COMPT" id"1"gt
• ltnodo tipo"SN" id"2"gt
• ltfoglia lemma"un" href"mw_002"/gt
• ltfoglia lemma"libro" href"mw_003"/gt
• ltnodo tipo"SA" id"3"gt
• ltfoglia lemma"bianco"
  href"mw_004"/gt
• lt/nodogt

•  
• ltnodo tipo"SPD" id"4"gt
• ltfoglia lemma"di"
  href"mw_005"/gt
• ltnodo tipo"SN" id"5"gt
• ltfoglia lemma"governo"
  href"mw_006"/gt
• ltnodo tipo"SN" id"6"gt
• ltfoglia lemma"major"
  href"mw_007"/gt
• lt/nodogt
• lt/nodogt
• lt/nodogt
• lt/nodogt
• lt/nodogt
• lt/nodogt
• ltfoglia lemma"." href"mw_008"/gt
• lt/nodogt
• lt/frasegt

13
Actual Work and Next Steps

• Precise the Meta Model for Syntactic Annotation
• Establish a list of data categories for syntactic
  annotation. Create a profile for this in the Data
  Category Registry at Loria.

14
WP3 Activities, Meetings, Synergies

• LIRICS WPs Meetings
• CNR-ILC DFKI, 5.5.2005 convergences between
  morpho-syntactic and syntactic data issues for
  the submission of the N W I on Syntax (SynAF) to
  ISO
• DFKI Tilburg, 2 meetings in Nov. And Dec. 2005
  on the relation betwenn argument structures and
  semantic roles.
• LORIA, CNR-ILC, DFKI, March 2006. Meeting in
  Paris discussing the data categories (with
  definition) to be included in the DCR. Redesign
  of DC, introduction of a profile for syntactic
  data categories.
• LIRICS Meetings
• Paris, 16-17.3.2005. Progress of work within WP3.
  Barcelona, 21-22.6.2005. LIRICS Industrial
  Advisory Board Meeting
• Barcelona, 22.6.2005 Presentation of first bulk
  of information relevant for lexical description
• Nancy, 8-9.12.2005. WP4 TDG3 Workshop
  convergences between lexico-semantic
  representation and semantic roles in lexicon and
  in annotation
• ISO Meetings
• Berlin 8-9.4.2005. ISO TC37/SC4 WG4 Meetings
• Warsaw 21-26.08.05. Plenary meeting of ISO
  TC37/SC4. Presentation of SynAF proposal (which
  has been accepted as NWI in July 2005)
• ISO TC37/SC4 meeting in Jeju, Korea (January
  2006). Discussion of SynAF with Asian and North
  American representative of standardisation bodies
• Joint ISO TC37/SC4, TDG3 and LIRICS meeting in
  Marina Del Rey discussing with American partners
  the issue of semantic annotation, incuding the
  relation between syntactic and semantic
  annotation