Bringing Computer Vision Training to Non Data Scientists in Academia

atit patel
2 min readMay 1, 2019

It would be an understatement to say that students today need to develop intuition around AI as part of their curriculum. We don’t want these students to graduate in next few years to find out that their core skills are being automated, do we? Akin to businesses, academic institutes have to be Agile and figure out ways to rapidly introduce AI into their core curriculum.

Until recently, you would have been forgiven if you believed that doing Computer Vision training required you to have deep data-science skills. One key challenge for developing an AI model is that a data scientist typically has very limited knowledge of the target domain. For example, you would not expect a Data Science major to understand how to read MRI scans. The opposite is also true in that it would be unreasonable to expect a medical student to build AI models that would classify/categorize/identify zones of interest in an MRI scan. So how then would an academic institute enable her students to learn AI hands-on as part of their curriculum? The Design and Development team at IBM grappled with this exact problem before developing PowerAI Vision software. PowerAI Vision software is designed to enable domain experts to guide the development of AI models without having deep data-science skills. PowerAI Vision has already received a thumbs-up from Mr. David…

--

--