Introduction

Today’s Day 1 of learning Machine Learning. So, I picked up a course called “Introduction to TinyML,” to start learning ML. This course is a little bit special, compared to your other ML courses. TinyML is about Machine Learning on Embedded Systems (tiny devices like a mobile phone, Arduino devices, and even Elon Musk’s NeuraLink!) And I went through the course introduction part, and it is really interesting!

Why This?

Now, I have several reasons for picking this course as my ML. They are:

• I always wanted to try and learn Embedded Systems too. And this one feels like a package deal for me.

• I want to learn ML fast (bad idea, I know) and the instructor on the course said that we won’t be getting all too deep in theory, just enough to understand ML. This is great because I can learn a little bit and then challenge me to solve problems other than the ones I was taught, and see if I’m able to use the same principles to explore, learn on my own, and solve for a different set of problems.

• There happens to be a lot of interesting things in this field and is being actively promoted by Harvard and they are supported by companies like Google, Arduino, and EdgeAI Foundation which again is supported by Dell, Sony, Qualcomm! So, something’s definitely cooking here!

I know I might sound like I’m getting sidetracked from Machine Learning here but there is definitely something cool happening with this and I just couldn’t resist from learn more about it.

I also found a bunch of resources like projects built by Harvard students who took this course as part of their academics, talks by experts in the field, and a few others.

You can find all of those links at mlnotes.dev/resources

About The Course(s)

So, this course is one out of the 3 courses. Here’s the full list:

1. Introduction to TinyML: Intro to ML with TensorFlow using Google Colab. Understand how to design, develop and use ML applications, especially for TinyML stuff.

2. Applications of TinyML: Intro to different TinyML applications, sensor types, and how to build some off these applications. Also includes about the importance of dataset engineering and responsible AI methods.

3. Deploying TinyML: Learn how to deploy models on a real microcontroller (an Arduino Nano) and explore several ML challenges that are amplified by TinyML (processing, post-processing, dealing with resource constraints, etc.).

From the way I see it, there’s just enough theory, inspiration, and then we jump right into building things. I think this is a solid structure for anyone who wants get into builder mode fast, without compromising on learning theory.

You can search “TinyML“ on edx.org to find these courses (and i also put a link in the resources page too).

I’m really excited for what I’ll be learning and finding next. Stay tuned if you’re curious too!