OPEN HOUSE on Multi-Task and Complex Outputs Learning

Multi-task feature learning

author: Andreas Argyriou, University College London

Description

We present a method for learning a low-dimensional representation which is shared across a set of multiple related tasks. The method builds upon the well-known 1-norm regularization problem using a new regularizer which controls the number of learned features common for all the tasks. We show that this problem is equivalent to a convex optimization problem and develop an iterative algorithm for solving it.

Categories

Top: Computer Science: Machine Learning: Structured Output
Top: Computer Science: Machine Learning: Preprocessing

You might be experiencing some problems with Your Video player.

*Slides*
0:02	Multi-Task Feature Learning
1:28	Learning Multiple Tasks Simultaneously
1:51	Multi-Task Feature Learning
3:52	Sharing Features Across Tasks
6:25	Learning Paradigm
8:16	Weighting Features
8:55	Sharing Features Across Tasks
9:26	Sharing Features Across Tasks
12:05	(2; 1)-Norm
13:18	(2; 1)-Norm
15:20	Sharing Features Across Tasks
16:22	(2; 1)-Norm
16:41	(2; 1)-Norm
16:52	(2; 1)-Norm Regularization
19:29	L1 Regularization
20:10	Learning the Features
21:22	Convex Reformulation
22:34	Convex Reformulation (cont.)
23:50	Alternating Algorithm
27:20	Convex Reformulation (cont.)
27:50	Alternating Algorithm
31:16	Experiment 1 (toy data)
32:35	Experiment 1 (toy data)
35:56	Experiment 2 (real data)
36:52	Experiment 2 (real data)
38:16	Experiment 2 (real data)
38:52	Summary
39:45	Future Work
41:30	Convex Reformulation (cont.)
46:55	Alternating Algorithm
47:45	Convex Reformulation (cont.)
48:20	Regularization with the Trace Norm

Lecture rating

People found this lecture:
Worth seeing
because it is:
Valuable and informative
Well presented
Easily understandable
Acceptably recorded
You need to login to cast your vote.

Report a problem or upload files

If you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.