The 25th International Conference on Machine Learning (ICML 2008)

Bayesian Multiple Instance Learning: Automatic Feature Selection and Inductive Transfer

author: Vikas Raykar, University of Maryland

Description

We propose a novel Bayesian multiple instance learning algorithm. This algorithm automatically identifies the relevant feature subset, and utilizes inductive transfer when learning multiple (conceptually related) classifiers. Experimental results indicate that the proposed baseline MIL method is more accurate than previous MIL algorithms and selects a much smaller set of useful features. Inductive transfer further improves the accuracy of the classifier as compared to learning each task individually.

You might be experiencing some problems with Your Video player.

*Slides*
0:00	Bayesian Multiple Instance Learning: Automatic Feature Selection and Inductive Transfer
0:17	Outline of the Talk
0:34	Outline of the Talk - Multiple Instance Learning
0:35	Binary Classification - 1
0:45	Binary Classification - 2
0:53	Binary Classification - 3
0:58	Linear Binary Classifier - 1
1:05	Linear Binary Classifier - 2
1:15	Linear Binary Classifier - 3
1:21	Linear Binary Classifier - 4
1:30	Labels for theTraining Data - 1
1:41	Labels for theTraining Data - 2
1:50	Single ’vs’ Multiple Instance Learning
2:00	MIL Applications
2:10	Computer Aided Diagnosis as a MIL Problem - 1
3:30	Computer Aided Diagnosis as a MIL Problem - 2
3:42	Our Notion of Bags - 1
3:52	Our Notion of Bags - 2
4:08	Our Notion of Bags - 3
4:14	MIL Illustration
4:24	Outline of the Talk - Proposed Algorithm
4:28	Proposed Algorithm
5:07	Training Data - 1
5:13	Training Data - 2
5:18	Training Data - 3
5:21	Training Data - 4
5:35	Classifier Form - 1
5:41	Classifier Form - 2
5:42	Single Instance Model
6:07	Multiple Instance Model - 1
6:38	Multiple Instance Model - 2
6:58	Maximum Likelihood (ML) Estimator - 1
7:05	Maximum Likelihood (ML) Estimator - 2
7:15	MAP Estimator - 1
7:21	MAP Estimator - 2
7:29	Our Prior
7:50	The Final MAP Estimator - 1
7:56	The Final MAP Estimator - 2
8:20	Our Prior
8:30	The Final MAP Estimator - 1
8:32	The Final MAP Estimator - 2
8:38	Outline of the Talk - Feature Selection
8:42	Feature Selection - 1
8:55	Feature Selection - 2
9:01	Feature Selection - 3
9:11	Feature Selection - 4
9:21	Feature Selection - 5
9:29	Feature Selection - 6
9:39	Feature Selection - 7
9:46	Feature Selection - 8
9:54	Feature Selection - 9
10:00	Feature Selection - 10
10:18	Feature Selection - 11
10:23	Feature Selection - 12
10:29	Feature Selection - 13
10:58	Outline of the Talk - Experiments
11:11	Benchmark Experiments
11:26	Experiments - 1
11:51	Experiments - 2
12:05	AUC Comparison
12:43	ROC Comparison - 1
12:51	ROC Comparison - 2
12:58	ROC Comparison - 3
13:04	ROC Comparison - 4
13:12	Features Selected
14:04	PECAD Experiments
14:20	ROC Comparison - 4
14:22	PECAD Experiments
15:21	Outline of the Talk - Multi-task Learning
15:31	Multi-task Learning - 1
16:08	Multi-task Learning - 2
16:49	Multi-task Learning Experiments
17:01	- Questions

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.