concepts.language#

Data structures for language semantics and tools for parsing. Most of this package are designed for neuro-symbolic representations of natrual language semantics and grammar-based language learning.

See the following papers for more details:

The Neuro-Symbolic Concept Learner: Interpreting Scenes, Words, and Sentences From Natural Supervision (http://nscl.csail.mit.edu/)
Grammar-Based Grounded Lexicon Learning (https://g2l2.csail.mit.edu/)

Here’s a quick summary of the sub-modules:

concepts.language.ccg contains tools for parsing sentences with (purely symbolic) Combinatory Categorial Grammars (CCGs).

`ccg`	Tools for parsing sentences with (purely symbolic) Combinatory Categorial Grammars.
`gpt_parsing`
`gpt_vlm_query`
`neural_ccg`	Data structures and algorithms for Neural CCG.
`openai_utils`