Semantic Code Classification for Automated Machine Learning

TL;DR

This paper introduces a semantic code classification approach to control automated machine learning processes by using simple action sequences, and discusses solving this task on the NL2ML dataset.

Contribution

It proposes a novel semantic code classification task for controlling AutoML and explores methods to address it using the NL2ML dataset.

Findings

Defined a new semantic code classification task

Presented initial methods for solving the task

Discussed potential applications in AutoML control

Abstract

A range of applications for automatic machine learning need the generation process to be controllable. In this work, we propose a way to control the output via a sequence of simple actions, that are called semantic code classes. Finally, we present a semantic code classification task and discuss methods for solving this problem on the Natural Language to Machine Learning (NL2ML) dataset.