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
Structured Linear Models
Published on Feb 25, 20077281 Views
Over the last five years, we have been able to extend the theory of linear classifiers to structure prediction problems, combining the benefits of discriminative learning and of structured probabilist
Related categories
Chapter list
Structured Linear Models00:00
Goals01:10
Information Extraction03:18
Biomedical Examples05:06
Approach07:04
Annotation Tool08:20
Analyzing Text10:00
Structured Classification10:32
Challenges11:45
Structured Classification (a)11:51
Challenges (a)11:59
Analysis by Tagging13:33
Segmentation as Tagging14:50
Traditional Approaches15:43
Hidden Markov Model17:10
HMMs in IE19:05
Problems with HMMs19:58
Hidden Markov Model (a)20:03
Problems with HMMs (a)20:15
Generating Multiple Features20:56
Structured Linear Models21:31
Learning25:06
Structured Linear Models (a)25:17
Learning (a)25:32
Margin28:55
Losses30:04
Why?34:17
Probabilistic Version35:15
Features36:15
Probabilistic Version (a)36:36
Features (a)36:43
MALLET37:44
Evaluation39:05
Gene/Protein Results39:44
Variation Results41:50
Tagger42:32
Fable43:18
Technical Challenges45:28
Alternative: Online Training47:32
Online Maximum Margin48:55
Lists and Unlabeled Text pt 251:45
Pattern Induction52:46
Person Names53:18
Improving CRF Tagger53:53
Extensions55:13
Losses (a)01:03:05
Learning (b)01:05:13