How AutoML Systems Accelerate Data Science

Automated machine learning (AutoML) is the process of automating the process of applying machine learning to real-world problems, with the stated aim of reducing or eliminating the need for skilled data scientists.

But with its ability to take care of the complete pipeline from raw dataset to deployable machine learning models – it automatically furnishes the labeled training data as input and outputs the optimized model. So does an AutoML system negate the need for skilled data scientists at the helm?

Automate with AutoML

In an article published on Forbes, Ryohei Fujimaki, the founder and CEO of dotData argues that the conversation is misplaced if the focus on AutoML systems is on replacing or reducing the role of the data scientist.

After all, the longest and most challenging aspect of a typical data science workflow revolves around feature engineering. This entails connecting data sources against a list of desired “features” that are evaluated against various Machine Learning algorithms.

Success with feature engineering requires a high level of domain expertise to identify the desired features through a time-consuming iterative process. Automation on this front allows even “citizen” data scientists to create optimized use cases by leveraging their domain expertise.

In a nutshell, this democratization of the data science process opens the door to new classes of developers, offering businesses a competitive advantage with minimum investments.

Of course, Fujimaki went on to advocate for the deployment of cutting-edge AutoML 2.0 platforms for its ability to accelerate and automate the process of discovering and creating features. This will allow for a more diverse and abundant group of users to contribute to the data science process, he notes.

And data scientists? Far from replacing them, the new tools let them address one of the most significant obstacles in their work. Specifically, they can leverage AutoML 2.0 platforms to dramatically accelerate their work, and more easily explore “unknown unknowns”.

Photo credit: iStockphoto/IvelinRadkov