Principles of Very Large Scale Modeling thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Principles of Very Large Scale Modeling

Published on Oct 07, 20147295 Views

ACM SIGKDD is pleased to announce that Pedro Domingos is the winner of its 2014 Innovation Award. He is recognized for his foundational research in data stream analysis, cost-sensitive classification,

Related categories

Chapter list

Principles of Very Large Scale Modeling00:00
Students and Collaborators00:23
Thanks from all of us00:44
IN MEMORIAM01:01
Road Map01:12
Very Large Scale Modeling01:51
The LSI-VLSI Transition03:14
A Similar Transition Is Underway in KDD05:40
A Similar Transition Is Underway in KDD08:27
First Principle08:56
Example: Social Networks09:02
Example: Social Networks09:33
Example: Social Networks10:01
Example: Social Networks10:11
Modeling the Whole Network10:35
Markov Logic Networks11:48
From Log-Linear Models to MLNs13:35
MLN for Viral Marketing15:16
Richer MLNs16:25
The Inference Problem17:50
The Cost of Inference19:37
The Cost of Inference20:25
The Cost of Inference20:47
The Cost of Inference21:24
Second Principle21:30
Hierarchical Decomposition21:54
The World Is Hierarchical23:33
The World Is Hierarchical24:17
Hierarchical Decomposition24:41
Exploiting Hierarchy25:16
Buying an Item27:24
Buying an Item28:22
Buying an Item29:19
Buying an Item29:50
The Sum-Product Theorem30:23
The Sum-Product Theorem32:15
Markov Logic Networks33:22
Third Principle34:00
From Big Data to Big Models34:21
Streaming Bound Algorithms35:16
Why Is This Possible?36:33
Learning an MLN38:02
Hoeffding Bounds40:23
Gradient Descent42:23
Gradient Descent Error Bound43:55
Gradient Descent Error Bound47:00
Streaming Gradient Descent48:09
What If Data Changes Over Time?50:58
Applications to Date51:47
Principles of Very Large Scale Modeling52:38
The Master Algorithm53:29