0.25
0.5
0.75
1.25
1.5
1.75
2
Learning and inference in the presence of corrupted inputs
Published on Aug 20, 20151900 Views
We consider a model where given an uncorrupted input an adversary can corrupt it to one out of $m$ corrupted inputs. We model the classification and inference problems as a zero-sum game between a lea
Related categories
Chapter list
Learning and inference in the presence of corrupted inputs00:00
Talk Overview00:06
Spam Detection00:28
Robust Spam Detection00:51
Network Failure01:56
Robust Network Failure Detection02:10
Adaptive model - 103:36
Adaptive model - 204:31
Adaptive model - 306:22
Our Main Results: Inference07:38
Our Main Results: Learning09:55
Where does it make a difference?12:02
Adaptive Adversary13:40
Modeling the setting13:49
Optimal policies - 114:34
Optimal policies - 215:21
Local Computation Algorithms - 115:46
Local Computation Algorithms - 216:14
From Matching to Vertex Cover16:51
Conclusion17:54
Thank you18:40