/schematizer

A sequence to sequence machine learning model to detect specific language patterns.

Primary LanguagePythonMIT LicenseMIT

schematizer

A sequence to sequence machine learning model to detect specific language patterns.

What is schematizer?

Schematizer is a machine learning model that transforms comments, comming from a comment-oriented programming language, into specific language patterns similar to regular expressions. The output of the model is then used by the schema parser generator.

But how?

To do so, a sequence to sequence RNN is used.

What does it look like?

let say we have the following comment:

let final_sum be the sum of a1 and a2.

The model should output the following sequence

let $var be the $op of $var and $var.