Generating Component-based Supervised Learning Programs From Crowdsourced Examples

We present CrowdLearn, a new system that processes an existing corpus of crowdsourced machine learning programs to learn how to generate e�ective pipelines for solving supervised machine learning problems. CrowdLearn uses a probabilistic model of program likelihood, conditioned on the current sequence of pipeline components and on the characteristicsof the inputdata to thenextcomponent in thepipeline, to predict candidate pipelines. Our results highlight the e�ectivenessof this technique in leveragingexistingcrowdsourced programs to generate pipelines that work well on a range of supervised learning problems.