DRASO: Declaratively Regularized Alternating Structural Optimization

author: Partha Pratim Talukdar, Computer & Information Science Department, University of Pennsylvania

Description

Recent work has shown that Alternating Structural Optimization (ASO) can improve supervised learners by learning feature representations from unlabeled data. However, there is no natural way to include prior knowledge about features into this frame- work. In this paper, we present Declar- atively Regularized Alternating Structural Optimization (DRASO), a principled way for injecting prior knowledge into the ASO framework. We also provide some analysis of the representations learned by our method.

You might be experiencing some problems with Your Video player.

*Slides*
0:00	DRASO: Declaratively Regularized Alternating Structural Optimization
0:41	Learning in Text and Language Processing
1:34	Alternating Structural Optimization (ASO)
2:10	Supervised Training in ASO
4:01	How does ASO work?
6:30	Auxiliary Problems for Sentiment Classification
7:58	Step 2: Training Auxiliary Problems
8:50	Using Prior Knowledge in ASO
10:46	Feature Similarity as Prior Knowledge
11:42	Model Feature Similarities with a Feature Graph
12:41	Regularization Criteria
13:59	Regularization in Auxiliary Problem Training
15:02	What Is the Effect of this New Regularizer?
16:58	Experimental Results
17:49	Comparing Learned Projections
20:08	Conclusion
21:01	Comparing Learned Projections
21:31	- Questions
21:32	- 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.