OPEN HOUSE on Multi-Task and Complex Outputs Learning

Going beyond bag-of-words: dealing with a text as a graph of triples

author: Marko Grobelnik, Department for Knowledge Technologies, Jožef Stefan Institute

Categories

Top: Computer Science: Machine Learning: Human Language Technology
Top: Computer Science: Information Extraction
Top: Computer Science: Machine Learning: Structured Output

You might be experiencing some problems with Your Video player.

*Slides*
0:00	Hierarchical Classification Into DMoz.Org
0:01	Hierarchical Classification Into DMoz.Org
0:27	Classification into large taxonomies
7:36	Max level:15 Avg level: 7
8:32	TITLE
8:39	TITLE
8:48	TITLE
8:54	TITLE
9:02	TITLE
9:11	TITLE
9:53	Max level:15 Avg level: 7
10:06	How our system works?
21:06	Query via URL
21:22	TITLE
22:15	Query via text
22:25	TITLE
23:38	Example classification of URL of a web page
23:58	Example classification of URL + text of a web page
24:08	Ontology Learning Challenge
36:00	Document summarization using semantic graphs
36:16	What is summarization?
37:08	Selection based summarization
37:10	slide21
37:32	Knowledge rich summarization
38:05	Our approach to summarization
38:42	Detailed Summarization procedure
40:06	Summarization: Steps 1 and 2
40:45	Linguistic analysis
41:02	Summarization: Steps 3 and 4
41:15	Named entities consolidation (1)
41:40	Named entities consolidation (2)
42:57	Anaphora resolution (1)
45:19	Anaphora resolution (2)
46:43	Anaphora resolution evaluation
47:03	Anaphora resolution evaluation
48:29	Summarization: Steps 5 to 7
48:34	Extracting triples
48:53	Constructing semantic graph (1)
49:22	WordNet
49:26	Constructing semantic graph (2)
50:11	Experiments
50:56	DUC 2001 dataset
50:59	Describing triples with attributes
53:09	Performance for various attribute sets
57:41	Insights
58:40	Example of automatic summary
58:51	Automatically generated graph of summary triples
58:58	Example of automatic summary
58:59	slide50
59:03	More examples
59:56	TITLE
60:15	Conclusion

Lecture rating

People found this lecture:
Worth seeing
because it is:
Valuable and informative
Well presented
Easily understandable
Acceptably recorded
You need to login to cast your vote.

Report a problem or upload files

If you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.