Previous Lecture Complete and continue  

  An interview with your instructor

Q: Hello Nimish, so I know the whole course is aimed at explaining this, but what is machine learning? Could you give us a brief run-through?

Nimish: Certainly; machine learning is basically a way for a program to analyze previous data (or past experiences) and use this knowledge to make decisions in the present or predict what might happen in the future.

There are many different approaches, but typically we train a model by showing it what outcomes it should expect given certain inputs so that it gets a sense for what a correct answer is and tries to find a pattern.

Q: Wow, that sounds pretty complex! But aren’t you claiming everyone can do it?

Nimish: Absolutely, and I stand by that claim! I won’t deny that everything going on behind the scenes is complex, including tight, efficient code and matrix algebra. However, we never need to touch that or even know about it, as we use frameworks (the main one we will use is called TensorFlow) that make it really easy to build, train, test and use machine learning models.

All you need to know is a little Python, which we will teach you, of course. TensorFlow has really been a saving grace, making machine learning so much more accessible to programmers everywhere.

Q: That’s good to know. I was under the impression that you had to be some sort of math whiz or have a formal education in data science to be able to build machine learning models.

So why did you decide to make this course now?

Nimish: Well, although machine learning frameworks like TensorFlow have been available for some time, it’s the recent advances that have made machine learning for Android and iOS devices feasible. Until fairly recently, mobile devices were just not powerful enough to execute computational heavy tasks such as those with machine learning.

However, mobile devices have become many times more powerful, and TensorFlow has recently come out with the mobile-optimized version TensorFlow mobile. And very recently, an even better version: TensorFlow Lite.

Once I read about TensorFlow Lite and realized that there are practically no courses out there on this topic, I decided to do something to change that. As I specialize in mobile app development, the advances that TensorFlow Lite will bring about really became the driving factor for me.

Q: That makes sense. Now, you said that there are almost no courses on the topic, but there are other courses on machine learning in general. How is your course different from the others?

Nimish: You’re right, there are plenty of courses on machine learning. However, there are next to no courses on big platforms like Kickstarter that focus on mobile machine learning in particular. All of them focus specifically on machine learning for a desktop or laptop environment.

As well, I’ve found that most courses don’t do a great job of explaining exactly what is going on at each step in the process and why we choose to build models the way we do. I aim to provide clear, concise explanations at each step along the way so that viewers can not only replicate, but also understand and expand upon what I teach.

Q: Right; there does seem to be a gap in the market for mobile machine learning courses. Final question: What can people expect from this Kickstarter?

Nimish: You can expect a complete and comprehensive course that guides you first through the basics, then through some simple models which we incorporate into apps. You will end up with a portfolio of apps driven by machine learning, as well as the know-how to create more and expand upon what we build together.

I start by teaching you the basics of the languages, programs and underlying concepts of machine learning. We then learn about subsets of the TensorFlow framework and how to build, train and test models and build Android and iOS apps around them.

Together we will build 10 or more apps. You will become an expert ready to build your own machine learning-driven mobile apps, which are the future in mobile app development.