Learn about transition-based parsing in natural language processing through this comprehensive lecture from CMU's Neural Networks for NLP course. Explore the fundamentals of shift-reduce parsing with feed-forward networks, stack LSTMs, and transition-based models for phrase structure. Discover the concept of linearized trees as a simple alternative approach. Gain insights into various parsing techniques, including arc standard shift-reduce parsing, recursive neural networks, and methods for encoding parsing configurations. Understand the importance of tree structures in NLP and how to make classification decisions in parsing tasks. Delve into advanced topics such as non-linear functions and alternative transition methods to enhance your understanding of neural network-based parsing techniques.
Overview
Syllabus
Intro
Two Types of Linguistic Structure
Why Dependencies?
Arc Standard Shift-Reduce Parsing (Yamada & Matsumoto 2003, Nivre 2003)
Shift Reduce Example
Classification for Shift-reduce
Making Classification Decisions
Non-linear Function: Cube Function
Why Tree Structure?
Recursive Neural Networks (Socher et al. 2011)
Encoding Parsing Configurations w/ RNNS
Alternative Transition Methods
Shift-reduce Parsing for Phrase Structure (Sagae and Lavie 2006. Watanabe 2015) . Shift, reduce X (binary), unary-X (unary) where X is a label
A Simple Approximation: Linearized Trees (Vinyals et al. 2015)
Taught by
Graham Neubig
Related Courses
-
Neural Nets for NLP 2017 - Transition-Based Dependency Parsing
-
Neural Nets for NLP 2020 - Generating Trees Incrementally
-
Neural Nets for NLP 2017 - Parsing With Dynamic Programs
-
CMU Multilingual NLP - Dependency Parsing
-
Neural Nets for NLP 2019 - Document Level Models
-
Neural Nets for NLP 2017 - Neural Semantic Parsing
