en-de
en-es
en-fr
en-pt
en-sl
en
en-zh
0.25
0.5
0.75
1.25
1.5
1.75
2
Online Structure Learning for Markov Logic Networks
Published on Nov 29, 20113318 Views
Most existing learning methods for Markov Logic Networks (MLNs) use batch training, which becomes computationally expensive and eventually infeasible for large datasets with thousands of training exam
Related categories
Chapter list
Online Structure Learning for Markov Logic Networks00:00
Large-scale structured/relational learning00:00
Motivation00:24
Outline01:23
Background01:35
Markov Logic Networks (MLNs)01:36
Existing structure learning methods for MLNs02:36
OSL: Online Structure Learner for MLNs03:18
Online Structure Learner (OSL)03:27
Max-margin structure learning04:30
Relational pathfinding05:58
Relational pathfinding (cont.)07:03
Mode declarations07:45
Mode-guided relational pathfinding08:10
Mode-guided relational pathfinding (cont.) - 108:48
Mode-guided relational pathfinding (cont.) - 209:57
Mode-guided relational pathfinding (cont.) - 310:25
Generalizing paths to clauses10:38
L1 - regularized weight learning11:30
Experiment Evaluation12:05
Input MLNs12:31
Input MLNs (cont.)13:26
Systems compared14:06
Experimental setup14:42
Average F1 scores on CiteSeer15:34
Average training time on CiteSeer16:50
Some good clauses found by OSL on CiteSeer17:15
Summary17:54
Thank you!18:46