Overview
- IBM and H2O.ai have partnered up to enable easier, quicker, efficient and interpretable machine learning tasks
- They have combined the IBM’s Power9 processor and H2O.ai’s Driverless AI
- The published results are incredible; feature engineering ran 1.5 times quicker than usual, time series models were developed at 5 times the usual speed
Introduction
IBM’s reach and power is a well known thing in the machine learning community. To enhance their reach into organizations and the wider ML community, they are now partnering up with H2O.ai.
In case you haven’t heard of H2O.ai, they are the developers of the ultra-popular Driverless AI platform. It is one of the best platforms for automated machine learning and has seen a rapid adoption rate since it’s launch. You don’t need to be a hardcore data scientist to enable your organization with ML skills anymore, tools like Driverless AI do the heavy lifting for you.
Under this partnership, IBM’s Power AI framework will work together with Driverless AI. This combined data science platform will aim to address a wide variety of real-life use cases for machine learning and deep learning in diverse industries. Power AI is basically an enterprise package of open source deep learning frameworks like PyTorch, Keras, TensorFlow, etc. It makes working with these packages easier and more efficient and reduces the training time significantly.
The combined platform has so far delivered incredible results. Driverless AI was run on Power9, IBM’s incredible fast and powerful processor. According to a blog post by H2O.ai, “Driverless AI is built on top of datatable for python for data ingest and feature engineering and H2O4GPU for machine learning”. With Power9, the team got almost 2 times the usual speed for data ingesting. The results were 50% faster for automated feature engineering.
The below table shows us how quickly analysis was performed for various tasks, using this platform:
Our take on this
This is a truly powerful collaboration between 2 industry leaders. I have previously used Driverless AI and can vouch for it’s incredible accuracy and ease of use. It doesn’t require coding, or a deep knowledge of data science (though obviously having that helps). Combining that with the ultra-fast Power9 will enable organizations to scale up their data science departments.
Both companies claim that it can be used in various industries which should see a rapid adoption rate once it’s publicly available.
Subscribe to AVBytes here to get regular data science, machine learning and AI updates in your inbox!
Senior Editor at Analytics Vidhya.Data visualization practitioner who loves reading and delving deeper into the data science and machine learning arts. Always looking for new ways to improve processes using ML and AI.
