Encoding Category Correlations into Bilingual Topic Modeling for Cross-Lingual Taxonomy Alignment thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Encoding Category Correlations into Bilingual Topic Modeling for Cross-Lingual Taxonomy Alignment

Published on Nov 28, 2017974 Views

Cross-lingual taxonomy alignment (CLTA) refers to mapping each category in the source taxonomy of one language onto a ranked list of most relevant categories in the target taxonomy of another language

Related categories

Chapter list

Encoding Category Correlations into Bilingual Topic Modeling for Cross-Lingual Taxonomy Alignment00:00
Outline00:27
Motivation00:39
Motivation - 100:59
Motivation - 201:29
Motivation - 302:18
Motivation - 403:10
Motivation - 503:55
Motivation - 604:11
Motivation - 704:42
Motivation - 805:20
Models – CC-BiLDA06:00
Models – CC-BiBTM06:59
Models – CC-BiBTM - 107:14
Models – Parameters Estimation07:38
Models – prior category distribution07:58
Models – prior category distribution - 108:44
Models – prior category distribution - 209:43
Models – prior category distribution - 310:32
Experiments10:54
Experiments - 111:37
Experiments – efficiency12:17
Conclusions12:49
Future work13:28