From Bandits to Experts : On the Value of More Information thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

From Bandits to Experts : On the Value of More Information

Published on Jul 25, 20113081 Views

Learning from Experts and Multi-armed Bandits are two of the most common settings studied in online learning. Whereas the first setting assumes that the performance of all k actions are revealed at th

Related categories

Chapter list

From Bandits to Experts: On the Value of More Information00:00
Experts / Multi-armed Bandits (1)00:04
Experts / Multi-armed Bandits (2)01:06
Model02:09
Examples (1)02:53
Examples (2)03:13
Examples (3)03:22
Motivation03:42
First attempt: the ExpBan Algorithm (1)05:01
First attempt: the ExpBan Algorithm (2)05:21
First attempt: the ExpBan Algorithm (3)05:27
First attempt: the ExpBan Algorithm (4)05:48
First attempt: the ExpBan Algorithm (5)06:17
Lower Bound07:37
Proof Intuition08:26
A Better Algorithm09:02
Regret (1)09:51
Proof Ideas10:54
Regret (2)11:49
Experiments12:37
Conclusions13:52
arXiv Tech Report14:47