Mining Significant Maximum Cardinalities in Knowledge Bases thumbnail
Pause
Mute
Subtitles not available
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Mining Significant Maximum Cardinalities in Knowledge Bases

Published on Dec 10, 201937 Views

Related categories

Chapter list

Mining Significant Maximum Cardinalities in Knowledge Bases00:00
Presentation Outline00:15
Using Web Knowledge Bases00:43
Related Works01:28
A Look on Role Cardinalities in DBpedia02:43
Take into account Incorrect Facts03:25
Another Look on Role Cardinalities in DBpedia03:38
Take into account Incompleteness04:02
More on Role Cardinalities in DBpedia04:19
Closer Look on Role Cardinalitiesin DBpedia04:52
Take into account Context and Distribution05:02
Computing True Maximum Cardinalities05:30
Assumptions: Level of Incorrectness06:15
Assumptions: Degree of Completeness06:38
Likelihood of a Maximum Cardinality i07:11
Examples of Likelihood Values08:02
Problem08:19
Correction Using Hoeffding’s Inequality08:34
Examples of Pessimistic Likelihood Values09:17
Significant Maximum Cardinality w.r.t. k09:34
Context10:14
Minimality10:51
C3M Algorithm11:28
On the Web Knowledge Base Scale…11:44
First Pruning Criteria: Significant11:53
Second Pruning Criteria: Minimal12:42
Experiments13:04
Good Scalability13:31
Manageable and interesting constraints13:58
High Precision14:29
Conclusion14:55
Merci15:46