Deep Learning for Efficient Discriminative Parsing thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Deep Learning for Efficient Discriminative Parsing

Published on May 06, 201117021 Views

We propose a new fast purely discriminative algorithm for natural language parsing, based on a "deep" recurrent convolutional graph transformer network (GTN). Assuming a decomposition of a parse tre

Related categories

Chapter list

Deep Learning for Efficient Discriminative Parsing00:00
Motivation (1)00:22
Motivation (2)01:32
Standard Parsing Benchmarks02:26
Parsing as a Tagging Task04:12
Neural Networks07:29
Words into Vectors08:09
Window Approach09:10
Window Approach (extra features)09:57
Sentence Approach10:11
Sentence Scoring11:36
Training13:14
Constrained Graph: Leverage BIOES15:09
Comparison With Pure Discriminative Parsers17:01
Comparison With State-Of-The-Art Parsers18:25
SENNA Demo19:19