0.25
0.5
0.75
1.25
1.5
1.75
2
An Alternating Direction Method for Dual MAP LP Relaxation
Published on Oct 03, 20113275 Views
Maximum a-posteriori (MAP) estimation is an important task in many applications of probabilistic graphical models. Although finding an exact solution is generally intractable, approximations based o
Related categories
Chapter list
An Alternating Direction Method for Dual MAP LP Relaxation00:00
The MAP Problem (1)00:00
The MAP Problem (2)00:17
MAP for Protein Design (1)00:33
MAP for Protein Design (2)00:41
The MAP Problem (3)00:51
Our Work02:06
The MAP Problem (4)02:49
The MAP Problem (5)03:36
MAP and LP Relaxation (1)04:09
MAP and LP Relaxation (2)05:34
MAP and LP Relaxation (3)06:11
MAP and LP Relaxation (4)06:24
The Dual LP06:26
Dual MAP-LP (1)07:56
Dual MAP-LP (2)08:31
Globally Optimal Methods09:04
Augmented Lagrangians (1)09:52
Augmented Lagrangians (2)10:42
Augmented Lagrangians (3)11:35
Augmented Lagrangians (4)11:38
Augmented Lagrangians (5)11:44
Augmented Lagrangians (6)11:59
Decomposable Objectives (1)12:18
Decomposable Objectives (2)13:43
Decomposable Objectives (3)13:48
ADMM for MAP-LP14:59
The Augmented Dual LP Algorithm (1)15:30
The Augmented Dual LP Algorithm (2)15:59
The Augmented Dual LP Algorithm (3)16:01
The Augmented Dual LP Algorithm (4)16:16
Updates (1)17:43
Updates (2)17:45
Updates (3)17:58
Updates (4)18:13
Updates (5)18:36
ADLP Algorithm18:44
Experiments (1)18:50
Experiments (2)19:31
Experiments (3)20:17
Experiments (4)20:29
Summary20:45
Thank you!21:02