University of Cambridge > > NLIP Seminar Series > Compositional Neural Meaning Representation Parsing

Compositional Neural Meaning Representation Parsing

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Guy Aglionby.

Much recent research on meaning representation parsing has focused on learning to map natural language sentences to rich semantic graphs, such as Dependency Minimal Recursion Semantics and Abstract Meaning Representation. In this talk, I will discuss composition-based methods for semantic graph parsing — a research effort in combining ‘old school’ knowledge-intensive models and neural data-intensive models. I will first introduce Context-free Graph Rewriting and show how to define symbolic grammars that map between strings and graphs in a linguistically-sound way. I will then discuss how to build a practical parsing system by integrating such a grammar with neural encoders and neural scorers. Finally, I will report on a parser which achieves state-of-the-art performance for Dependency Minimal Recursion Semantics. The graph rewriting framework also supports mapping semantic graphs to surface strings.

Join Zoom Meeting

Meeting ID: 917 8973 9002 Passcode: 155692

This talk is part of the NLIP Seminar Series series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.


© 2006-2024, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity