en
0.25
0.5
0.75
1.25
1.5
1.75
2
Selective Sampling for Information Extraction with a Committee of Classifiers
Published on 2007-02-254529 Views
Related categories
Presentation
Selective Sampling for Information Extraction with a Committee of Classifiers00:00
Overview00:21
Approaches to Active Learning00:56
Committee02:08
Feature Split03:24
KL-divergence (McCallum & Nigam, 1998)04:15
Evaluation Results05:26
Discussion06:08
Other Selection Metrics06:41
Costing Active Learning08:23
Related Work: BioNER09:14
Simulation Results: Sentences09:51
Simulation Results: Tokens10:20
Simulation Results: Entities10:40
Costing AL Revisited (BioNLP data)10:53
Document Cost Metric (Dev)12:31
Token Cost Metric (Dev)12:48
Longest Document Baseline13:46
Confusion Matrix (1)13:48
Confusion Matrix (2)14:36
Confusion Matrix (3)16:17
Confusion Matrix (4)16:38
Conclusions18:00
Thank you19:10