SLIDE 1
Automatically Extracting new Search Patterns for VerbOcean
Kuzman Ganchev
1
Overview
- VerbOcean refresher
- How patterns are used
- What makes a good pattern
- How we look for patterns
- Results
2
Automatically Extracting new Search Patterns for VerbOcean Kuzman - - PowerPoint PPT Presentation
Automatically Extracting new Search Patterns for VerbOcean Kuzman - - PowerPoint PPT Presentation
Automatically Extracting new Search Patterns for VerbOcean Kuzman Ganchev 1 Overview VerbOcean refresher How patterns are used What makes a good pattern How we look for patterns Results 2 VerbOcean Refresher: relations
SLIDE 2
SLIDE 3
VerbOcean Refresher: relations
Find relations between verbs
Relation Examples Synonymous Stronger persuade→advise
- verhaul→amend
Similar absolve→vindicate allot→earmark Opposite absolve→vindicate allot→earmark Causes centralize→overhaul privatize→improve Happens Before advise→persuade survive→avenge 3
SLIDE 4
VerbOcean Refresher: patterns
Search google for verb pairs in “patterns”.
Relation Pattern Examples Synonymous “X ie Y” “Xed ie Yed” Stronger “X even Y” “Yed or at least Xed” Similar “Xed and Yed” “X and Y” Opposite “either Xed or Yed” “to X * but Y” Causes “Xed by Ying the” “Xed by Ying or” Happens Before “to X and then Y” “Xed and later Yed” “Xed and later Yed” + “buy → sell” = “bought and later sold” Pattern Verb Pair 4
SLIDE 5
How Patterns are Used
- p
Sp(V1, V2) > C where: Sp(V1, V2) =
hits(V1,p,V2) N hitsest(p) N
× hits(”toV1”)×Cv
N
× hits(”toV2”)×Cv
N
for asymmetric relations and Sp(V1, V2) =
hits(V1,p,V2) N hits(V2,p,V1) N
2 × hitsest(p)
N
× hits(”toV1”)×Cv
N
× hits(”toV2”)×Cv
N
for symmetric relations.
5
SLIDE 6
What Makes a Good Pattern
For asymmetric relations: arg max
p
Sp(V1, V2) = hits(V1, p, V2) hitsest(p) For symmetric relations: arg max
p
Sp(V1, V2) = hits(V1, p, V2) × hits(V2, p, V1) hitsest(p) .
6
SLIDE 7
How we look for patterns
- List of verb pairs in known relations
- Wildcard Search: “buy * * sell”, “bought * * sold”
- Extract patterns: “X and then Y”, “Xed and later
Yed”
- Rank patterns.
7
SLIDE 8
Ranking Patterns: ideas
- find arg maxp
- V1,V2 Sp(V1, v2)
- use the odds ratio: arg maxp
P
V1,V2∈R hits(V1,p,V2)
P
V1,V2∈R hits(V1,p,V2)
8
SLIDE 9
Training data
100 verb-pairs (used in testing VerbOcean).
Relation Occurances Examples Synonymous Stronger 19 persuade→advise
- verhaul→amend
analyse→monitor devastate→batter doom→complicate erode→impact Similar 26 absolve→vindicate allot→earmark avenge→survive bellow→scream bobble→throw commute→reverse Opposite 6 absolve→vindicate allot→earmark avenge→survive bellow→scream bobble→throw commute→reverse Causes 2 centralize→overhaul privatize→improve Happens Before 12 advise→persuade survive→avenge batter→devastate grab→carry commence→suspend decree→reconsider
9
SLIDE 10
Results
10
SLIDE 11
Comments
- Need lots of verb pairs for this to be viable.
- Restricting contribution of each verb pair.
- Getting wildcard patterns like “Xed * by Ying the”
11