In this talk, I summarize the research conducted during my visiting period at the Web&Media Group of the Vrije Universiteit, Amsterdam.
Extracting relevant entities in TV-programs descriptions is a challenging problem, due to the broad amount of topics they cover and the different formats they have. None of existing tools for automatic Named-Entity Recognition and Classification is trained with these data.
I illustrate the workflow established for extracting relevant entities from a text in the entertainment domain,
relying on the adoption of different annotators, as well as the issues arising in
the integration of their outputs. In order to increase the coverage of the annotation task,
metrics based on majority-vote are combined with metrics established for the crowd-truth evaluation for gold-standard creation. This approach should be able of capturing cases
typically cut off by majority-vote integration techniques (i.e., unique information and distributed agreement).
Several features are computed in order to capture as many characteristics as possible, useful for assessing the relevance of an entity. Human annotators results, gathered through a crowd-sourcing task, are used for collecting positive and negative examples of relevance and, as an ultimate goal, for evaluating precision and recall of the entire system.
Abortion pills in Dammam Saudi Arabia// +966572737505 // buy cytotec
What is the relevant information in a text?
1. What is the relevant information in a text?
Silvia Giannini
Visiting PhD student
Politecnico di Bari
Web & Media Group meeting | 27.10.2014
2. The scenario
•Entertainment domain: BBC TV-programs (TV-series, movies, documentaries, …)
•Aim: Enrich the content description with links to the Web of Data
•Applications: Linked Data patterns for recommendations; multi- domain datasets creation, …
Following the grandeur of Baroque, Rococo art is often dismissed as frivolous and unserious, but Waldemar Januszczak disagrees. The first episode is about travel in the 18th century and how it impacted greatly on some of the finest art ever made. The world was getting smaller and took on new influences shown in the glorious Bavarian pilgrimage architecture, Canaletto's romantic Venice and the blossoming of exotic designs and tastes all over Europe. The Rococo was art expressing itself in new, exciting ways.
3. How?
•Yet another semantic annotation tool?
•Peculiarities:
- Different formats
- Broad coverage of topics
Following the grandeur of Baroque, Rococo art is often dismissed as frivolous and unserious, but Waldemar Januszczak disagrees. The first episode is about travel in the 18th century and how it impacted greatly on some of the finest art ever made. The world was getting smaller and took on new influences shown in the glorious Bavarian pilgrimage architecture, Canaletto's romantic Venice and the blossoming of exotic designs and tastes all over Europe. The Rococo was art expressing itself in new, exciting ways.
4. Multiple annotators integration
Following the grandeur of Baroque, Rococo art is often dismissed as frivolous and unserious, but Waldemar Januszczak disagrees. The first episode is about travel in the 18th century and how it impacted greatly on some of the finest art ever made. The world was getting smaller and took on new influences shown in the glorious Bavarian pilgrimage architecture, Canaletto's romantic Venice and the blossoming of exotic designs and tastes all over Europe. The Rococo was art expressing itself in new, exciting ways.
text enrichment
“Canaletto”
ontology:Location
“Rococo”
dbpedia:Rococo_(band)
•Type mis-classification
•URI mis-annotation
•Not relevant labels
The NERD framework
5. Multiple annotators integration
•Feature-based solution for entity relevance definition and entity classification
•Majority vote and disagreement metrics
•Extractors can disagree on:
–The existence of a label, e.g. some can identify a lable and other can’t
–The span of the label, e.g. ‘Myra’ VS ‘Myra Gail’
–The type of the label, e.g. ‘Building’ VS ‘Organization’
–The URI of the label
Proposal
6. Multiple annotators integration
DEFINE THE RELEVANCE OF A LABEL
•Feature-based solution for entity relevance definition and entity classification
•Majority vote and disagreement metrics
•Extractors can disagree on:
–The existence of a label, e.g. some can identify a lable and other can’t
–The span of the label, e.g. ‘Myra’ VS ‘Myra Gail’
–The type of the label, e.g. ‘Building’ VS ‘Organization’
–The URI of the label
Proposal
7. Multiple annotators integration
INFLUENCE THE CONFIDENCE OF AN ANNOTATION
•Feature-based solution for entity relevance definition and entity classification
•Majority vote and disagreement metrics
•Extractors can disagree on:
–The existence of a label, e.g. some can identify a lable and other can’t
–The span of the label, e.g. ‘Myra’ VS ‘Myra Gail’
–The type of the label, e.g. ‘Building’ VS ‘Organization’
–The URI of the label
Proposal
8. Workflow for relevance assessment
Following the grandeur of Baroque, Rococo art is often dismissed as frivolous and unserious, but Waldemar Januszczak disagrees. […] The first episode is about travel in the 18th century and how it impacted greatly on some of the finest art ever made. The world was getting smaller and took on new influences shown in the glorious Bavarian pilgrimage architecture, Canaletto's romantic Venice and the blossoming of exotic designs and tastes all over Europe. The Rococo was art expressing itself in new, exciting ways.
Relevant labels
crowd
NLP tools
metrics
Features matrix extraction
Classifier
9. •Disagreement between TextRazor annotations in NERD and standalone TextRazor in terms of missing labels, missing types, granularity of types.
PID: b0074t2b
Title: Great plains
Synopsis
‘’The great plains are the vast open spaces of our planet. […] Close on their heels come an array of plains predators including eagles, wolves and lions. […]‘’
Label
eagles#439#445
Extractors
Types URI
textrazor(nerd)
nerd:Thing http://en.wikipedia.org/wiki/Eagle
textrazor
dbpedia-owl:Bird http://en.wikipedia.org/wiki/Eagle
Workflow for relevance assessment
10. Pre-processing
•Alignment of extractors’ results:
-Label: each label has a list of alternative labels contained in or overlapping with the given one
-Type: same vocabulary for all extraction methods (529 classes of the Dbpedia ontology, extended with owl:Thing and Amount type)
-URI: Dbpedia resources
•Label
•NERD ontology class
•sameAs link
•Label
•DBpedia ontology class
•Wikipedia page
•Label
•DBpedia category
•Wikipedia page
•Label
•DBpedia ontology class
•DBpedia URI
11. Majority-vote for relevance: longest-span strategy*
extractor
label
startOffset
endOffset
Aligned label
Rococo
35
41
Rococo art#35#45
Rococo art
35
45
Rococo art#35#45
Rococo Art
35
42
41
45
Rococo art#35#45
Rococo art#35#45
Rococo Art
35
42
41
45
Rococo art#35#45
Rococo art#35#45
•Label
•NERD ontology class
•sameAs link
•Label
•DBpedia ontology class
•Wikipedia page
•Label
•DBpedia category
•Wikipedia page
•Label
•DBpedia ontology class
•DBpedia URI
Label & span alignment
The LONGEST-SPAN strategy
*Analogously, the shortest-span strategy can be applied
12. Issues
•In the previous example, Rococo and Art are related to the same category (Arts). Thus, the longest-span strategy for labels alignment will lead to a consistent conceptual category for the new label (Rococo Art).
•Consider this program description:
A journey back to the 1950s for a look at the wildest pop music of all time in a film that tells the stories of Bill Haley, Elvis Presley, Little Richard, Chuck Berry, Jerry Lee Lewis and Buddy Holly, giants from an era when pop music really was mad, bad and dangerous to know.The programme features the artists themselves, alongside people like Bill Haley's original Comets, the Crickets, Buddy Holly's widow Maria Elena, Jerry Lee Lewis's former wife Myra Gail and his sister, Chuck Berry's son and many more, including June Juanico, Elvis' first serious girlfriend.Other contributors include Tom Jones, Jamie Callum, Paul McCartney, Cliff Richard, Joe Brown, Marty Wilde, Green Day, Minnie Driver, Jack White, the Mavericks, Jools Holland, Hank Marvin, Fontella Bass, John Waters and more.Elvis's pelvis was just the start. Who had to change the lyrics to their biggest hit because the originals were too obscene? Who married their 13-year-old cousin? Who used lard to get their hair just right? And what happened on the day the music died?
BBC Program: Kings of Rock and Roll (Pid: b007c95q)
13. Issues
•In the previous example, Rococo and Art refer to the same conceptual category. Thus, the longest-span strategy for labels alignment will lead to a consistent conceptual category for the new label (Rococo Art).
•Consider this program description:
BBC Program: Kings of Rock and Roll (Pid: b007c95q)
extractor
label
startOffset
endOffset
Type
Aligned label
Myra Gail
453
462
Person
Myra Gail#453#462
Myra Myra Gail
453
453
457
462
Settlement
Person
Myra Gail#453#462
Myra Gail#453#462
Myra
Gail
453
458
457
462
Band,Artist
Person
Myra Gail#453#462
Myra Gail#453#462
Myra Gail
453
462
Thing
Myra Gail#453#462
14. The HYBRID-SPAN strategy1
Given two labels l1 and l2 and an upper ontology O, l1 and l2 belong to the same annotation span if:
1. l1 is contained in l2 or l2 is contained in l1 and type(l1) and type(l2) are in super(sub)class relationship (e.g. Royal Academy[Organization] in Royal Academy of Music[University])
OR
2. l1 and l2 are overlapping but neither l1 is contained in l2 nor l2 is contained in l1 (e.g., Royal Academy[Organization] and Academy of Music[Building])
OR
3. l1 coincides with l2 (e.g., Royal Academy[Organization] and Royal Academy[Museum])
What about Thing type?
1Chen, L., Ortona, S., Orsi, G., & Benedikt, M. (2013). Aggregating Semantic Annotators. Proceedings of the VLDB Endowment, Vol. 6, No. 13, (p. 1486-1497). Riva del Garda, Trento, Italy.
15. •Label
•NERD ontology class
•sameAs link
•Label
•DBpedia ontology class
•Wikipedia page
•Label
•DBpedia category
•Wikipedia page
•Label
•DBpedia ontology class
•DBpedia URI
Label & span
alignment
The HYBRID-SPAN strategy*
extractor
label
startOffset
endOffset
Type
Aligned label
Myra Gail
453
462
Person
Myra Gail#453#462
Myra
Myra Gail
453
453
457
462
Settlement
Person
Myra#453#457
Myra Gail#453#462
Myra
Gail
453
458
457
462
Band,Artist
Person
Myra#453#457
Myra Gail#453#462
Myra Gail
453
462
Thing
Myra Gail#453#462
*The vocabulary alignment is required as previous step
Majority-vote for relevance: hybrid-span strategy
16. Features for Relevance
•F1: nerd(l) -> 1 if label l is extracted by NERD;
0 otherwise
label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
17. Features for Relevance
•F2: textrazor(l) -> 1 if label l is extracted by TextRazor;
0 otherwise
label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
18. Features for Relevance
•F3: tagme(l) -> 1 if label l is extracted by TAGME;
0 otherwise
label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
19. Features for Relevance
•F4: nltk(l) -> 1 if label l is extracted by the NLTK-based method;
0 otherwise
label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
20. Features for Relevance
•F5: abs(l) = 푛푒푟푑푙+푡푒푥푡푟푎푧표푟푙+푡푎푔푚푒푙+푛푙푡푘푙 |퐸푀|
Absolute score for l over the set EM of all Extraction Methods (four in this setting)
label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
21. Features for Relevance
label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
•F6: lss(l) = 푤푐푙 푤푐(푙LS) , where wc is the word count function and lLS is the longest span containing l in the union set of all labels recognized by each extraction methods
Expresses the span overlapping between l and the longest span containing l, i.e. the portion of l contained in the longest span lLS
22. Features for Relevance
•F7: wlss(l) = 푎푏푠푙∗푙푠푠(푙)
Longest span score for l, weighted by the absolute score for l
label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
23. Features for Relevance
•F8: sss(l) = 푤푐푙SS 푤푐(푙) , where lSS is the shortest span contained in l in the union set of all labels recognized by each extraction methods
Expresses the span overlapping between l and the shortest span contained in l, i.e. the portion of l containing the shortest span lSS
label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
24. Features for Relevance
•F9: wsss(l) = 푎푏푠푙∗푠푠푠(푙)
Shortest-span score for l, weighted by the absolute score for l
label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
25. Features for Relevance
•F10: oss(l) = |푗 ∩푙| |푗 ∪푙|푗 ∈푂퐿 |푂퐿| , where |OL| is the number of overlapping labels among the alternative ones.
label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
26. Features for Relevance
•F11: woss(l) = 표푠푠푙∗푎푏푠(푙)
label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
27. Features for Relevance with type
Label#offset
type
Alternative label
F1
…
wildlife
#912#920
Thing
1
east africa
#361#372
Thing
africa
#366#372
[Place,Continent]
0
Country
0
africa
#366#372
Place
east africa
#361#372
[Thing,Country]
1
Continent
1
…
28. Features for Relevance with type
•F1: nerd(l,t) -> 1 if label l with type t is extracted by NERD; 0 otherwise
•F2: textrazor(l,t) -> 1 if label l with type t is extracted by TextRazor; 0 otherwise
•F3: tagme(l,t) -> 1 if label l with type t is extracted by TAGME; 0 otherwise
•F4: nltk(l,t) -> 1 if label l with type t is extracted by the NLTK-based method; 0 otherwise
29. Features for Relevance with type
Label#offset
type
Alternative label
F1
F2
F3
F4
F5a
F5b
F6
…
east africa
#361#372
Thing
africa
#366#372
[Place,Continent]
0
1
0
0
0.25
0.33
0.5
Country
0
0
1
1
0.5
0.67
0.5
•F5a: abs(l,t) = 푛푒푟푑푙,푡+푡푒푥푡푟푎푧표푟푙,푡+푡푎푔푚푒푙,푡+푛푙푡푘푙,푡 |퐸푀|
Absolute score for l with type t over the set EM of all Extraction Methods (four in this setting)
•F5b: rel(l,t) = 푎푏푠푙,푡 푎푏푠푙
Relative score for label l with type t over the total number of extraction methods recognizing l
30. Features for Relevance with type
•F6: lss(l,t) = 푙푠푠(푙) 푛_푐푎푡(푙) , where n_cat is the number of different types associated with l.
Expresses the span overlapping between l and the longest span containing l, weighted by the number of different types associated with the same label l.
Label#offset
type
Alternative label
F1
F2
F3
F4
F5a
F5b
F6
…
east africa
#361#372
Thing
africa
#366#372
[Place,Continent]
0
1
0
0
0.25
0.33
0.5
Country
0
0
1
1
0.5
0.67
0.5
31. Features for Relevance with type
•F7a: wlss(l,t) = 푎푏푠푙,푡∗푙푠푠(푙,푡)
Longest span score for l with type t, weighted by the absolute score for label l and type t
•F7b: wrlss(l,t) = 푟푒푙푙,푡∗푙푠푠(푙,푡)
Longest span score for l with type t, weighted by the relative score for label l and type t
Label#offset
type
Alternative label
F1
F2
F3
F4
F5a
F5b
F6
…
east africa
#361#372
Thing
africa
#366#372
[Place,Continent]
0
1
0
0
0.25
0.33
0.5
Country
0
0
1
1
0.5
0.67
0.5
32. Features for Relevance with type
•F8: sss(l,t) = 푠푠푠(푙) 푛_푐푎푡(푙)
•F9a: wsss(l,t) = 푎푏푠푙,푡∗푠푠푠(푙,푡)
•F9b: wrsss(l,t) = 푟푒푙푙,푡∗푠푠푠(푙,푡)
Label#offset
type
Alternative label
F1
F2
F3
F4
F5a
F5b
F6
…
east africa #361#372
Thing
africa
#366#372
[Place,Continent]
0
1
0
0
0.25
0.33
0.5
Country
0
0
1
1
0.5
0.67
0.5
33. Features for Relevance with type
•F10: oss(l,t) = 표푠푠(푙) 푛_푐푎푡(푙)
•F11a: woss(l,t) = 푎푏푠푙,푡∗표푠푠(푙,푡)
•F11b: wross(l,t) = 푟푒푙푙,푡∗표푠푠(푙,푡)
Label#offset
type
Alternative label
F1
F2
F3
F4
F5a
F5b
F6
…
east africa
#361#372
Thing
africa
#366#372
[Place,Continent]
0
1
0
0
0.25
0.33
0.5
Country
0
0
1
1
0.5
0.67
0.5
34. Features for Relevance with type
Label#offset
type
Alternative label
F1
F2
F3
F4
…
F12
F13
…
east africa #361#372
Thing
africa
#366#372
[Place,Continent]
0
1
0
0
1
0.375
Country
0
0
1
1
0.5
0.17
•F12: hss(l,t) = |푖푛푇푟푒푒퐴퐿푙,푡| |퐴퐿| , where |inTreeAL(l,t)| is the number of Alternative Labels in the set AL with type in a sub(super)-sumption relation with t
•F13: whss(l,t) = 1 푑푡푙,푡푗+1/|퐴퐿|푗 ∈푖푛푇푟푒푒퐴퐿, where
|d(tl, tj)| is the distance between class tl and tj in the ontology
35. •Disagreement on the extractors corner (i.e., tools that more sistematically disagree with every other tool) could reveal:
- bad quality tools (in recognizing specific set of labels/types)
- specialized tools able to recognized particular entities better than all the other tools
Disagreement metrics evaluation
on the extractors corner2
Disagreement for relevance:
Humans VS Machine Annotation
2G. Soberon, L. Aroyo, C. Welty, O. Inel, H. Lin, M. Overmeen, Measuring Crowd Truth: Disagreement Metrics Combined with Worker Behavior Filters, Proc. of CrowdSem2013 Workshop, ISWC2013.
36. Features for Relevance
Label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
•DISTRIBUTED AGREEMENT
•UNIQUE INFORMATION
37. Features for Relevance
Label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
•ela(ei, ej, l) = 풆풊풍∗풆풋풍 |푳(풆풊,풑)| , where 푖≠푗. 푒푖푙 is the corresponding extractor score (F1-4) and 푳풆풊,풑 the number of labels recognized by extractor i in program p (the extractor-label agreement operator is not commutative)
38. Features for Relevance
Label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
•avg_ela(ei, l) = 풆풍풂(풊≠풋풆풊,풆풋,풍) |푬푴|
Average extractor-label agreement over the set of extraction methods
39. Features for Relevance
Label#offset
Alternative labels
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
wildlife
#912#920
1
1
0
1
0.75
1
0.75
1
0.75
0
0
east africa
#361#372
africa
#366#372
0
1
1
1
0.75
1
0.75
0.5
0.375
0
0
africa
#366#372
east africa
#361#372
1
1
0
0
0.5
0.5
0.25
1
0.5
0
0
earth: two
#227#237
two million
#234#245; earth
#227#232; two million gazelles
#234#254
0
0
1
0
0.25
1
0.25
0.5
0.125
0.29
0.07
…
•Both extractor-label agreement and the consequent average are evaluated also with reference to the pairs (label,type)
40. Other possible relevance features
•TF-IDF (with type)
Shall the corpus for idf contain more episodes of the same TV-series?
Labels referring to characters mentioned in many episodes of the same TV series will gain a higher tf but lower idf score -> consider metadata
Animated adventures of Pingu, the clumsy young penguin. Pingu helps his neighbour and is rewarded. Pingu's friend tries to get a reward too, but the neighbour refuses. They decide to play a trick on the neighbour, but it all ends with an innocent passer-by becoming the victim of their prank.
BBC Program: Pingu's Trick (Pid: b0077x84)
41. •Enhance metadata (words in title and subject)
Labels lemmatization (WordNetLemmatizer)
Dani is understudying the part of a witch in Macbeth: The Musical, which means Jack and Sam get the job of ensuring little brother Max does not cause chaos. Dani's most loyal viewers, the aliens, have got bored of never getting to meet their heroine and her pals, and have decided to teleport down to Earth, where they soon find themselves embroiled in Max's scheme to win the 10,000 pound reward from the UFO Society.
BBC Program: Alien Invasion (Pid: b00ph91v)
Other possible relevance features
42. State of work
•Dataset: 52 BBC programs
•Realized:
- Span and Type Alignment
- Relevance scores for labels
•To do:
–Computation of relevance score for pairs (label,type)
–Crowdsourcing tasks
–Connecting relevance/relevance-with-type outputs
–Evaluation of results (precision, recall, complementarity, …)
43. Does the method deal with complementarity?
http://dbpedia.org/resource/Gazelle
PID: b0074t2b
Title: Great plains
Synopsis
‘’The great plains are the vast open spaces of our planet. These immense wilderness areas are seemingly empty. But any feeling of emptiness is an illusion - the plains of our planet support the greatest gatherings of wildlife on earth: two million gazelles on the Mongolian steppes, three million caribou in North America and one and a half million wildebeest in East Africa. […]‘’
Label
two million gazelles#234#254
Types
Amount;Mammal;Single
Extractors
wikimeta(nerd);textrazor;tagme;
http://dbpedia.org/resource/Two_in_a_Million/You're_My_Number_One
COMPLEMENTARITY!!
(Amount of Mammal)