91 - (Executable) Semantic Parsing, with Jonathan Berant
NLP Highlights
00:00
How to Build a System That Translates From Language to an Executable Representation
The model is like there's a grammar part, a structural part, and a statistical part. And your goal is to learn a model that will assign high probabilities for the correct translation given the input. Because the search space is very large, this is a non-trivial problem. So you usually need also a parsers or some kind of learning algorithm.
Transcript
Play full episode
