Tense and Implicit Role Reference

1
Tense and Implicit Role Reference

• Joel Tetreault
• University of Rochester
• Department of Computer Science

2
Implicit Role Reference

• Verb phrases have certain required roles NPs
  that are expected
• For example take
• Something to take (theme)
• A place to take it from (from-loc)
• A place to take it to (to-loc)
• Something to do the taking (agent)
• Possibly a tool to do the taking (instrument)

• Very little work has been has been done
• (Poesio, 1994 Asher and Lascarides, 1998)

3
Goal

• Resolving IRRs important to NLP
• To investigate how implicit roles work
• Develop an algorithm for resolving them
• Evaluation of algorithm for empirical results
• Use temporal information and discourse relations
  to improve results

4
Outline

• Implicit Roles
• Annotation
• Algorithm
• Results
• Discussion

5
Example

• (1) Take engine E1 from Avon to Dansville
• (2a) Pick up the boxcar and take it to Broxburn
  from ?
• (2b) And then take the boxcar from Corning
• to ?
• (3a) Leave E1 there but move the boxcar down the
  road to Evansville. from ?
• (3b) Leave the boxcar there.

6
Statistics Role Distributions
7
Corpus

• Annotated a subset of the TRAINS-93 Corpus
  (Heeman and Allen, 1994)
• 86 utterance task-oriented dialog between two
  humans
• Task move commodities around in a virtual world

8
Annotation

• Used sgml style annotation scheme
• NPs annotated with ID and its class (engine,
  tanker, location, food)
• VPs annotated with ID, event time, and roles
• Roles for each verb are taken from TRIPS lexicon
  (Allen et al., 2000)

9
Temporal Annotation

• An event time was assigned each utterance, such
  as t0, t1, u1, etc.
• And constraints upon the event time are imposed
• t9t1 (t9 comes after t1)
• t9
• t9t1 t9

10
Sample Annotation

• U1 Take Engine E1 from Avon to Dansville.
• U2 Pick up the boxcar
• to-locne6Take engine E1 from
  Avon to Dansville/ve.
• t0 themene13
  from-locne6iPick upthe
  boxcar.

11
Statistics

• Most implicit roles have antecedents found
  locally (0-2 sentences back over 90 of the time)
• Instrument 79 Instr, 10 theme, 10 ID
• Theme 88 Theme, 12 ID
• From-Loc 62 From-Loc, 38 To-Loc
• To-Loc 57 To-Loc, 29 From-Loc, 14 Theme

12
Algorithm

• For each utterance u, process u left to right
• If NP is encountered, push it on appropriate
  focus stack
• If VP is encountered
• place all explicit roles on top of appropriate
  focus stack
• If role is implicit.

13
Algorithm Example
U1 Take engine E1 from Avon to Dansville
Avon
Dansville
Engine E1
empty
Theme
From-Loc
To-Loc
Instrument
U2 Also take the boxcar
Avon
Dansville
boxcar
empty
Theme
From-Loc
To-Loc
Instrument
14
Implicit Role Algorithm

• Type determines method. If role is
• Instrument search through current utterance
  first for an entity that meets verbs
  constraints, else go back through past
  utterances instrument and theme focus lists
• Theme same as above except search theme before
  instrument for each past utterance
• From/To-Loc use temporal reasoning to determine
  what order to search past To-Loc and From-Loc
  lists for each utterance

15
Temporal Algorithm

• For two utterances uk and uj,with k j,
  determine rel(uk, uj)
• If time(uk) time(uj) then rel(uk, uj)
  narrative
• Else rel(uk, uj) parallel

16
Experiment

• Developed LISP system that automates the
  algorithm
• For each marked implicit role, system tries to
  find an antecedent
• Notations
• R-L each focus list searched right to left
  (order of recency)
• L-R search is left-to-right (sentence order)
• Time algorithm augmented with temporal algorithm

17
Results
18
Discussion

• From-Locs naive version is better
• To-Loc any strategy better than naïve
• Top pronoun resolution algorithms perform around
  70-80 accuracy
• Problems
• Corpus size hard to make concrete conclusions
  or find trends
• Annotation scheme is basic
• Need to handle return verbs properly
• Augment system to identify whether implicit roles
  should be resolved or not (ignore general cases)

19
Current Work

• Building a larger corpus that can be annotated
  automatically using the TRIPS parser
• Domain is much more varied and has different
  types of verbs