Large Scale Learning Which Is Actually Useful

author: Ronan Collobert, NEC Laboratories America, Inc.

You might be experiencing some problems with Your Video player.

*Slides*
0:00	Large Scale Learning Which Is Actually Useful
2:16	The Goal (1)
2:43	The Goal (2)
3:16	Natural Language Processing
5:27	How Large-Scale Is It By The Way?
6:05	SVMs with 1M of Labeled Examples
9:18	SVMs with ∞ Unlabeled Examples
10:58	Large Scale = Complex Models (1)
11:36	Large Scale = Complex Models (2)
12:54	Large Scale = Complex Models (3)
14:11	NLP: Large Scale Engineering (1/2)
15:05	NLP: Large Scale Engineering (2/2)
16:03	NLP: Large Scale Learning (1/2)
17:29	NLP: Large Scale Learning (2/2)
18:16	Neural Networks?!
19:40	The Deep Learning Way (1/2)
22:22	The Deep Learning Way (2/2)
23:29	Convolutions
24:23	The Deep Learning Way (2/2)
25:09	Convolutions
25:10	Removing The Time Dimension (1/2)
25:59	The Deep Learning Way (2/2)
26:31	Removing The Time Dimension (1/2)
26:33	Removing The Time Dimension (2/2)
26:35	Multi-Task Learning
27:27	The Deep Learning Way (2/2)
28:01	Multi-Task Learning
28:25	1M of Words is not Large Scale Enough!
30:25	Improving Word Embedding
32:08	Multi-Task Learning
32:21	Improving Word Embedding
32:27	Language Model: Think Massive
37:05	The Deep Learning Way (1/2)
38:01	Language Model: Think Massive
38:06	Common Pitfall
39:03	Child Learning
39:53	Language Model: Embedding
40:37	MTL: Semantic Role Labeling
41:49	MTL: Unified Network for NLP
42:21	Noodle - Description
43:06	Noodle - Experiments
43:19	Noodle - Demo
43:38	Noodle vs Google
44:03	Summary
45:10	Conclusion
47:05	- 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.