First ACM International Conference on Web Search and Data Mining - WSDM 2008

Machine Reading at Web Scale

author: Oren Etzioni, Turing Center, University of Washington

Categories

Top: Computer Science: Machine Learning: Semi-supervised Learning
Top: Computer Science: Natural Language Processing
Top: Computer Science: Web Mining

You might be experiencing some problems with Your Video player.

*Slides*
0:00	Machine Reading of Web Text
0:50	Rorschach Test
0:59	Rorschach Test for CS
1:08	Moore’s Law?
1:09	Storage Capacity?
1:10	Number of Web Pages?
1:12	Number of Facebook Users?
1:14	Turing Center
1:30	Turing Center Foci
4:51	Outline
5:10	Web Search in 2020?
6:20	Intelligent Search
7:56	Application: Information Fusion
10:24	Opinion Mining
11:12	Opine (1)
13:00	Opine (2)
13:44	But “Reading” the Web is Tough
15:18	Semi-Supervised Learning
16:20	2. Open IE = Self-supervised IE
17:22	Extractor Overview (Banko & Etzioni, ’08)
18:50	TextRunner Extraction
19:45	Numerous Extraction Challenges
20:49	TextRunner (Web’s 1stOpen IE system)
21:23	TextRunner Demo
21:40	TextRunner Search (1)
22:22	TextRunner Search (2)
23:48	Sample of 9 million Web Pages
24:58	3. Tractable Inference
26:08	I. Entity Resolution
27:26	Relation Synonymy
28:18	II. Probability of Correctness
29:23	Counting Extractions
31:00	Formal Problem Statement
31:53	Combinatorial Model (“Urns”)
33:24	Performance (15x Improvement)
34:20	URNS limited on 'sparse' facts
35:37	Language Models to the Rescue
37:46	III Compositional Inference (work in progress, Schoenmackers, Etzioni, Weld)
40:47	KnowItAll Family Tree
42:15	KnowItAll Team
42:31	Related Work
42:32	4. Conclusions
44:09	Thank you

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.