How classifieres can be use to solve any reasonable loss thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

How classifieres can be use to solve any reasonable loss

Published on Feb 25, 20073238 Views

Related categories

Chapter list

How Classifiers can be ...00:02
A Machine Learning Dream00:27
Two Versions of the Dream03:14
Basic Question04:57
The Reductionist Style for ...06:14
Defining the Learning Problem08:27
Step (1) Class Probability Estimation15:36
Step (2) The Probing Method: Observations17:00
The Probing Algorithm21:47
Step (2) The Probing Method: Detailes23:27
Step (3) The one classifier trick27:23
Step (3) Probing Theory32:11
What Are Learning Reductions?51:44
Disadvantages of Learning Reductions52:39
Error Limiting Reductions54:09
Uses of Error Limiting Reductions57:13
Regret Transform Reductions57:17
Uses of Regret Transform Reductions58:13
Big Questions58:18