DIY AI Has Arrived With Jetson Nano
By Cait Etherington
July 12, 2019
In the past DIY (do-it-yourself) was usually a scissors and glue or hammer and nails affair. Today, as artificial intelligence (AI) continues to transform how we work, learn, and live, AI is becoming increasingly accessible to non-experts and even beginning to transform our hobbies. eLearning Inside recently talked to Amit Goel, the Product Line Manager of Autonomous Machines at NVIDIA, about how Jetson Nano and NVIDIA’s Deep Learning Institute are turning AI into a DIY possibility for everyone, even people with no or a very limited existing knowledge of programming and deep learning.
Meet Jetson Nano
Cait Etherington – When most of us think about AI, we don’t think about DIY. Can you first tell us a bit about Jetson Nano? What is it and how does it make AI a DIY affair?
Amit Goel – That’s exactly the myth that we want to debunk with Jetson Nano. We started validating our belief that anyone can learn and deploy AI a few years back when we recruited high school interns who knew nothing about AI/deep learning and let them experiment with it.
The results were fantastic and exceeded our expectations. To bring AI to the DIY community, we had to solve two big challenges: accessibility and ease of use. We’ve addressed this with Jetson Nano and the response from the maker/DIY community has been overwhelming.
Jetson Nano is an AI computer that offers a full desktop Linux environment that most developers are familiar with and supports all popular programming languages, making it the easiest platform to start learning AI. Also, since most DIY projects involve controlling some actions, Jetson Nano offers superior performance to enable these real-world applications.
Types of DIY AI Projects
CE – What sorts of projects have people already developed using Jetson Nano? Do you see any trends emerging as regular folks start to tinker with AI solutions?
AG – There are a wide variety of Jetson Nano projects that people are developing ranging from autonomous RC cars to home automation. We have created a list of some of the projects on our site.
Scaling AI Training With the Deep Learning Institute
CE – You’ve just launched a Deep Learning Institute to help people get started with Jetson Nano. What background would you need to enroll in the course, and what can enrollees expect to learn? In other words, is the course for just anyone or only semi experts?
AG – The specific requirements are posted on the DLI course page. But as a quick overview, for anyone who wants to enroll, a basic familiarity with Python would be helpful but it’s not required. In the course, you’ll learn to set up your Jetson Nano and camera, collect image data for classification models, annotate image data for regression models, run interference, and much more. Upon completion, you’ll be able to create your own deep learning classification and regression models with the Jetson Nano.
A Future of AI Hobbyists
CE – As AI technologies become increasingly accessible, what’s next? What interesting innovations and even hobbies do you see developing out the emerging DIY AI field? And how can AI be embraced in a positive and generative way?
AG – AI is helping us solve some of the most complex problems of our generation. It’s making our transportation systems safer and more efficient, our farms more productive, and our medical professionals more equipped to provide better care for people. An interesting hobby that’s emerging out of the DIY AI field is the racing of autonomous toy cars. It’s a big AI trend that’s emerging–influenced by the growing interest in autonomous vehicles.
Image of Amit Goel courtesy of NVIDIA.