NIPS Workshop on Structured Input - Structured Output, Whistler 2008

Learning to Predict Combinatorial Structures

author: Shankar Vembu, Department of Computer Science, University of Illinois at Urbana-Champaign
published: Dec. 20, 2008, recorded: December 2008, views: 142

Slides

Slides
0:00	Learning to Predict Combinatorial Structures
0:05	Learning Structured Prediction Models (1)
0:35	Learning Structured Prediction Models (2)
0:45	Learning Structured Prediction Models (3)
0:56	Combinatorial Structures
1:15	Outline
1:48	Structured Prediction
2:10	Discriminative Structured Prediction
2:43	Assumptions
4:02	Optimality vs. Non-optimality
4:45	Dicycles - Optimality vs. Non-optimality (1)
5:28	Dicycles - Optimality vs. Non-optimality (2)
5:43	Optimality vs. Non-optimality
6:13	Training Combinatorial Structures
6:22	Loss Functions (1)
6:47	Loss Functions (2)
7:01	Loss Functions (3)
7:22	Regularised Risk Minimisation (1)
8:05	Regularised Risk Minimisation (2)
8:15	Optimisation with Finite Output Embedding
8:53	Recipe for Training Combinatorial Structures (1)
9:07	Recipe for Training Combinatorial Structures (2)
9:11	Recipe for Training Combinatorial Structures (3)
9:20	Recipe for Training Combinatorial Structures (4)
9:39	Constructing Combinatorial Structures
9:46	Approximation Algorithms
9:58	Approximation Measure
10:07	z-approximation
10:17	Decoding Sibling Systems (1)
10:50	Decoding Sibling Systems (2)
10:58	Decoding Independence Systems (1)
11:14	Decoding Independence Systems (2)
11:29	Application settings
11:35	Multiclass
12:39	Multilabel
13:10	Dicycles (1)
13:31	Dicycles (2)
14:24	And others
14:51	Experiments
14:54	Multilabel Classification
15:07	Hierarchical Classification
15:25	Dicycle Policy Estimation (1)
15:42	Dicycle Policy Estimation (2)
16:39	Ongoing Work
17:33	Questions
18:15	- Questions

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.

Lecture popularity: You need to login to cast your vote.