relational duality: unsupervised extraction of semantic relations between entities on the web...
TRANSCRIPT
![Page 1: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/1.jpg)
Relational Duality:Unsupervised Extraction of Semantic Relations
between Entities on the Web
Danushka BollegalaYutaka Matsuo
Mitsuru Ishizuka
International World Wide Web Conference 2010Rayleigh, North Carolina, USA
![Page 2: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/2.jpg)
Relation Extraction on the Web Problem definition
Given a crawled corpus of Web text, identify all the different semantic relations that exist between entities mentioned in the corpus.
Challenges The number or the types of the relations that exist in
the corpus are not known in advance Costly, if not impossible to create training data Entity name variants must be handled
Will Smith vs. William Smith vs. fresh prince,… Paraphrases of surface forms must be handled
acquired by, purchased by, bought by,… Multiple relations can exist between a single pair of
entities
![Page 3: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/3.jpg)
Relational Duality
ACQUISITION
(Microsoft, Powerset)
(Google, YouTube)
…
X acquires YX buys Y for $ …
Extensional definition Intensional definitionDUALITY
![Page 4: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/4.jpg)
Overview of the proposed method
Web crawler
Text Corpus
Sentence splitter
POS Tagger
NP chunker
Pattern extracto
r
Lexical patterns
Syntactic patterns
Entity pairs vs. Patterns Matrix
(Google, YouTube)(Microsoft, Powerset)
:::
X a
cquir
es
YX
buys
Y: :
Sequential Co-
clustering Algorithm
Entity pair clusters
Lexico-syntactic pattern clusters
Cluster labeler (L1 regularized
multi-class logistic
regression)
![Page 5: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/5.jpg)
Lexico-Syntactic Pattern Extraction Replace the two entities in a sentence by X
and Y Generate subsequences (over tokens and POS
tags) A subsequence must contain both X and Y The maximum length of a subsequence must be L
tokens A skip should not exceed g tokens Total number of tokens skipped must not exceed G Negation contractions are expanded and are not
skipped Example
… merger/NN is/VBZ software/NN maker/NN [Adobe/NNP System/NN] acquisition/NN of/IN [Macromedia/NNP]
X acquisition of Y, software maker X acquisition of Y X NN IN Y, NN NN X NN IN Y
![Page 6: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/6.jpg)
Entity pairs vs. lexico-syntactic pattern matrix Select the most frequent entity pairs and
patterns, and create an entity-pair vs. pattern matrix.
Entity pairs vs. Patterns Matrix
(Google, YouTube)(Microsoft, Powerset)
:::
X a
cquir
es
YX
buys
Y: :
![Page 7: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/7.jpg)
Sequential Co-clustering Algorithm1. Input: A data matrix, row and column clustering thresholds
2. Sort the rows and columns of the matrix in the descending order of their total frequencies.
3. for rows and columns do: Compute the similarity between current row (column) and
the existing row (column) clusters If maximum similarity < row (column) clustering threshold:
Create a new row (column) cluster with the current row (column)
else: Assign the current row (column) to the cluster with the maximum
similarity
repeat until all rows and columns are clustered
4. return row and column clusters
![Page 8: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/8.jpg)
Sequential Co-clustering Algorithm
0 5 0 2 1
0 8 0 3 2
5 0 1 1 0
6 0 8 2 0(Google, YouTube)
(Microsoft, Powerset)
(Balmer, Microsoft)
(Jobs, Apple)
X a
cquir
ed Y
Y C
EO
X
X b
uys
Y f
or
$
X o
f Y
Y h
ead X
=8
=13
=7
=16
Row clustering threshold = column clustering threshold = 0.5
![Page 9: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/9.jpg)
Sequential Co-clustering Algorithm
0 5 0 2 1
0 8 0 3 2
5 0 1 1 0
6 0 8 2 0(Google, YouTube)
(Microsoft, Powerset)
(Balmer, Microsoft)
(Jobs, Apple)
X a
cquir
ed Y
Y C
EO
X
X b
uys
Y f
or
$
X o
f Y
Y h
ead X
=11 =13 =9 =8 =3
![Page 10: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/10.jpg)
Sequential Co-clustering Algorithm
5 0 0 2 1
8 0 0 3 2
0 5 1 1 0
0 6 8 2 0(Google, YouTube)
(Microsoft, Powerset)
(Balmer, Microsoft)
(Jobs, Apple)
X a
cquir
ed Y
Y C
EO
X
X b
uys
Y f
or
$
X o
f Y
Y h
ead X
=11=13 =9 =8 =3
![Page 11: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/11.jpg)
Sequential Co-clustering Algorithm
5 0 0 2 1
8 0 0 3 2
0 5 1 1 0
0 6 8 2 0(Google, YouTube)
(Microsoft, Powerset)
(Balmer, Microsoft)
(Jobs, Apple)
X a
cquir
ed Y
Y C
EO
X
X b
uys
Y f
or
$
X o
f Y
Y h
ead X
![Page 12: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/12.jpg)
Sequential Co-clustering Algorithm
5 0 0 2 1
8 0 0 3 2
0 5 1 1 0
0 6 8 2 0[(Google, YouTube)]
(Microsoft, Powerset)
(Balmer, Microsoft)
(Jobs, Apple)
X a
cquir
ed Y
Y C
EO
X
X b
uys
Y f
or
$
X o
f Y
Y h
ead X
![Page 13: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/13.jpg)
Sequential Co-clustering Algorithm
5 0 0 2 1
8 0 0 3 2
0 5 1 1 0
0 6 8 2 0[(Google, YouTube)]
(Microsoft, Powerset)
(Balmer, Microsoft)
(Jobs, Apple)
X a
cquir
ed Y
[Y C
EO
X]
X b
uys
Y f
or
$
X o
f Y
Y h
ead X
0.067 < 0.5
![Page 14: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/14.jpg)
Sequential Co-clustering Algorithm
5 0 0 2 1
8 0 0 3 2
0 5 1 1 0
0 6 8 2 0[(Google, YouTube)]
(Microsoft, Powerset)
[(Balmer, Microsoft)]
(Jobs, Apple)
X a
cquir
ed Y
[Y C
EO
X]
X b
uys
Y f
or
$
X o
f Y
Y h
ead X
0<0.5
![Page 15: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/15.jpg)
Sequential Co-clustering Algorithm
5 0 0 2 1
8 0 0 3 2
0 5 1 1 0
0 6 8 2 0[(Google, YouTube)]
(Microsoft, Powerset)
[(Balmer, Microsoft)]
(Jobs, Apple)
[X a
cquir
ed Y
]
[Y C
EO
X]
X b
uys
Y f
or
$
X o
f Y
Y h
ead X
0.071
0.998 >0.5
![Page 16: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/16.jpg)
Sequential Co-clustering Algorithm
13 0 0 5 3
0 5 1 1 0
0 6 8 2 0[(Google, YouTube)]
(Microsoft, Powerset)
[(Balmer, Microsoft), (Jobs,Apple)]
[X a
cquir
ed Y
]
[Y C
EO
X]
X b
uys
Y f
or
$
X o
f Y
Y h
ead X
0
0.84 > 0.5
![Page 17: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/17.jpg)
Sequential Co-clustering Algorithm
13 0 5 3
0 6 1 0
0 14 2 0
[(Google, YouTube)]
(Microsoft, Powerset)
[(Balmer, Microsoft), (Jobs,Apple)]
[X a
cquir
ed Y
,X
bu
ys
Y f
or
$]
[Y C
EO
X]
X o
f Y
Y h
ead X
0.99 > 0.5
0.05
![Page 18: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/18.jpg)
Sequential Co-clustering Algorithm
13 0 5 3
0 20 3 0
[(Google, YouTube),(Microsoft, Powerset)]
[(Balmer, Microsoft), (Jobs,Apple)]
[X a
cquir
ed Y
,X
bu
ys
Y f
or
$]
[Y C
EO
X]
X o
f Y
Y h
ead X
0.85 > 0.5 0.51
![Page 19: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/19.jpg)
Sequential Co-clustering Algorithm
18 0 3
3 20 0
[(Google, YouTube),(Microsoft, Powerset)]
[(Balmer, Microsoft), (Jobs,Apple)]
[X a
cquir
ed Y
,X
bu
ys
Y f
or
$]
[Y C
EO
X,
X o
f Y]
Y h
ead X
0.98 > 0.5 0
![Page 20: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/20.jpg)
Sequential Co-clustering Algorithm
21 0
3 20
[(Google, YouTube),(Microsoft, Powerset)]
[(Balmer, Microsoft), (Jobs,Apple)]
[X a
cquir
ed Y
,X
bu
ys
Y f
or
$]
[Y C
EO
X,
X o
f Y,
Y h
ead X
]
Entity pair clusters
Lexical-syntactic pattern clusters
A greedy clustering algorithmAlternates between rows and columnsComplexity O(nlogn)Common relations are clustered firstThe no. of clusters is not requiredTwo thresholds to determine
![Page 21: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/21.jpg)
Estimating the Clustering Thresholds Ideally each cluster must represent a unique semantic
relation Number of clusters = Number of semantic relations Number of semantic relations is unknown Thresholds can be either estimated via cross-validation
(requires training data) OR approximated using the similarity distribution.
Similarity distribution is approximated using a Zeta distribution (Zipf’s law)Ideal clustering:
inter-cluster similarity = 0 intra-cluster similarity =meanwith a large number of data points:
average similarity in a cluster ≥ threshold threshold ≈ distribution mean
![Page 22: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/22.jpg)
Measuring Relational Similarity Empirically evaluate the clusters produced
Use the clusters to measure relational similarity (Bollegala, WWW 2009)
Distance = ENT dataset: 5 relation types, 100 instances Task: query using each entity pair and rank using relational
distanceRelation VSM LRA EUC RELSIM Proposed
ACQUSITION 0.92 0.92 0.91 0.94 0.89
HEADQUARTERS 0.84 0.82 0.79 0.86 0.97
FIELD 0.44 0.43 0.51 0.57 0.42
CEO 0.95 0.96 0.90 0.95 0.99
BIRTHPLACE 0.27 0.27 0.33 0.36 0.53
Overall Average Precision
0.68 0.68 0.69 0.74 0.76
![Page 23: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/23.jpg)
Entity pair clusters
Self-supervised Relation Detection What is the relation represented by a cluster?
Label each cluster with a lexical pattern selected from that cluster
C1 C2 Ck
Entity pairs vs. Patterns Matrix
(Google, YouTube)(Microsoft, Powerset)
:::
X a
cquir
es
YX
buys
Y: :
(Google,YouTube)=[X acquired Y:10,…]
Train an L1 regularized multi-class logistic regression Model (MaxEnt) to discriminate the k-classes.Select the highest weighted lexical patterns from each class
![Page 24: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/24.jpg)
Subjective Evaluation of Relation Labels Baseline
Select the most frequent lexical pattern in a cluster as its label
Ask three human judges to assign grades A: baseline is better B: proposed method is better C: both equally good D: both badRelation A B C D
ACQUSITION 16.7% 40% 40% 3.3%
HEADQAURTERS 20% 40% 23.3% 16.7%
CEO 6.7% 53.3% 20% 20%
FIELD 13.3% 56.7% 23.3% 6.7%
BIRTHPLACE 13.3% 36.7% 10% 40%
Overall 14% 45.3% 23.3% 17.3%
![Page 25: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/25.jpg)
Open Information Extraction Sent500 dataset (Banko and Etzioni, ACL
2008) 500 sentences, 4 relation types Lexical patterns 947, syntactic patterns 384 4 row clusters, 14 column clustersMethod Precision Recall F
O-NB 0.866 0.232 0.366
O-CRF 0.883 0.452 0.598
MLN 0.798 0.733 0.764
PROP (lexical) 0.943 0.647 0.767
PROP (syntactic) 0.752 0.860 0.802
PROP (lexical + syntactic) 0.751 0.857 0.801
![Page 26: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/26.jpg)
Classifying Relations in a Social Network
spysee.jp
![Page 27: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/27.jpg)
Relation Classification Dataset
790,042 nodes (people), 61,339,833 edges (relations)
Randomly select 50,000 edges and manually classify into 53 classes
11,193 lexical patterns, 383 pattern clusters, 664 entity pair clusters
Relation P R F Relation P R F
colleagues
0.76 0.87 0.81 friends 0.58 0.77 0.66
alumni 0.83 0.68 0.75 co-actors 0.75 0.74 0.74
fan 0.91 0.50 0.64 teacher 0.83 0.73 0.78
husband 0.89 0.57 0.74 wife 0.67 0.34 0.45
brother 0.79 0.60 0.68 sister 0.90 0.52 0.66
Micro 0.72 0.68 0.70 Macro 0.78 0.52 0.63
![Page 28: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/28.jpg)
Conclusions Dual representation of semantic relations leads
to a natural co-clustering algorithm. Clustering both entity pairs and lexico-syntactic
patterns simultaneously helps to overcome data sparseness in both dimensions.
Co-clustering algorithm scales nlog(n) with data Clusters produced can be used to:
Measure relational similarity with performance comparable to supervised approaches
Open Information Extraction Tasks Classify relations found in a social network.
![Page 29: Relational Duality: Unsupervised Extraction of Semantic Relations between Entities on the Web Danushka Bollegala Yutaka Matsuo Mitsuru Ishizuka International](https://reader036.vdocuments.mx/reader036/viewer/2022062809/5697bff11a28abf838cbb654/html5/thumbnails/29.jpg)
Thank You
29