Abstract

We introduce an automatic machine learning (AutoML) modeling architecture called Autostacker, which combines an innovative hierarchical stacking architecture and an Evolutionary Algorithm (EA) to perform efficient parameter search. Neither prior domain knowledge about the data nor feature preprocessing is needed. Using EA, Autostacker quickly evolves candidate pipelines with high predictive accuracy. These pipelines can be used as is or as a starting point for human experts to build on. Autostacker finds innovative combinations and structures of machine learning models, rather than selecting a single model and optimizing its hyperparameters. Compared with other AutoML systems on fifteen datasets, Autostacker achieves state-of-art or competitive performance both in terms of test accuracy and time cost.

https://arxiv.org/pdf/1803.00684.pdf

Boyuan Chen Columbia University bchen@cs.columbia.edu Harvey Wu Columbia University wu.harvey@columbia.edu Warren Mo University of Chicago warrenmo@uchicago.edu Ishanu Chattopadhyay University of Chicago ishanu@uchicago.edu Hod Lipson Columbia University hod.lipson@columbia.edu

This paper is really including two previous ideas  in mind, and telling me how to continue my AutoML work:

1. How to find the best combination of various models
2. How to take the advantage of the output of various models during training or even testing

The lesson I learned is that never release out your idea till you’re ready!

Lastly,  I would like to keep the source  (Note: In Chinese) the link containing many related AutoML works.