This time, our SPOTLIGHT interview throws 10 questions at the youngest superstar in the world of Machine Learning and A.I. - Tanmay Bakshi!
For those not yet familiar with Tanmay, his resume is very impressive. His LinkedIn headline describes some of his achievements, including; Keynote & TEDx speaker, Algorithmist, Author, Watson & Cognitive Developer, YouTuber
On that last point - Tanmay has over 133k subscribers to his 'Tanmay Teaches' channel - it's well worth a look!
- Tanmay delivers his TEDx talk
I'm sure you'll agree that his achievements in the field thus far are truly remarkable - now add in the fact that he was born in 2003...
Anyway, on with the interview - enjoy!
What was your first experience with a computer - and how did this experience start you on your journey to where you are today?
My very first experience with a computer, and in fact, programming, was when I was 5 years old. At the time, my dad worked as a computer programmer, and as you can imagine, watching him program was really fascinating to me, as a 5-year-old! Due to the fact that I started at such a young age, I was able to pick up programming as a hobby, and not as “work”, which I believe in to this day. Since this was so fun for me, I started to learn through numerous sources, such as books and the internet. This led me to learn iOS Development at 8, and release my first iOS app at 9. From that point onwards, I developed multiple other iOS Applications, and when I was 11 years old, and when I stumbled upon IBM Watson, I set myself on a new mission: to use AI and Machine Learning to the benefit of all humanity.
You've written a book on, and have a number of Videos relating to Swift - do you utilise other popular machine learning languages like Python & R?
Definitely! I love languages like Python, and in fact, I love to use Python for my general scripting and for architecting, training, and, testing my deep learning systems; and to use Swift to deploy my deep learning systems using CoreML on Apple platforms. Mainly, I use Keras as my library of choice, with TensorFlow as a backend generally, and CNTK as a backend for RNNs.
Who are your inspirations in the field of Machine Learning & AI - what is it that makes them so inspiring?
My inspirations in the field of AI would have to be people like Elon Musk and companies like Apple, because you can see how seriously they take Artificial Intelligence, and the rapid pace at which they’ve adopted it. In fact, you can see that Tesla was one of the very first, if not the first, car company to bring AI based intelligent self-driving to the masses. You can also see how Apple foresaw the advantages of AI, and even created Machine Learning-based systems that make it more efficient to store images! They’re also working to bring ML to the hands of more developers, by incorporating SDKs like CoreML into iOS, macOS, watchOS, and tvOS.
What does a typical day look like for you at the moment?
The typical day involves my school studies, around 40% goes to coding, and the rest goes to some recreation, depending on the weather. I could either choose biking, table tennis, or preparing and recording YouTube videos, or even helping out those who reach out to me 1-to-1. When I’m in a mood and have the ideas flowing I write blogs, articles, or work on my book’s chapters.
What are you currently working on?
A lot of people have the misconception that I work for IBM, but that’s not the case – I love the services that IBM Watson provides, and I love to use them in my applications. I’m currently working on a variety of projects that use Watson’s capabilities; including, but not limited to, facial recognition apps, spam detection and reporting systems, and systems that allow the disabled to communicate via AI. I’m also the host of an IBM Facebook Live series called ‘Watson Made Simple with Tanmay’. I’m an Honorary IBM Cloud advisor and Champion for cloud.
What do you see as the 'next big thing' in the field on Deep Learning & AI?
It’s really hard to tell, because this field is evolving so rapidly due to the fact that it’s relatively new (it was invented several decades ago, but became popular 5 years ago). However, I can see that in order for deep learning to become even bigger than it already is, it needs to be faster. Users don’t want to wait 5 seconds every time they want to talk to their chatbot – it becomes annoying. So, with massive advances in both computing power (AI ASICs, e.g. IBM TrueNorth, Google Tensor Processing Unit, NVIDIA Tensor Cores), and Deep Learning optimization techniques (Pruning, GAs), I can see Neural Nets becoming way bigger.
In your TEDx talk you say that "we should not be afraid of A.I" - what are your thoughts on Elon Musk's contrary views?
I believe that Elon Musk does have a point – but the fact that he’s arguing is something that I believe people misunderstand. AI is very capable of doing something negative, but only if it is instructed what to do and how to do so, by other humans. If an AI is trained to drive you around the city, it won’t turn against humanity and overpower us – its objective function is constant. Elon is arguing the point that humanity itself may use AI to its own disadvantage, which is completely valid itself, but I say this: this has been done, and is being done, with all technologies that have ever existed. We’re still combating against it, and we can do the same with AI.
What are your personal goals for the next 5 years - what would you like to achieve?
In the next 5 years, I’ll continue to work towards my goal of reaching out to 100,000 aspiring beginners, along their journey of innovating through coding, of which I’m 5,000 people there. I’d also like to continue to develop and improve Neural Network architectures, and performance. Of course, I also love to share my knowledge, and not only do I want to learn, but I want to share everything that I learn with the rest of the world, so we can bring the power and capabilities of Deep Learning into the hands of even more developers across the globe. Alongside, I’m most passionate about using my AI algorithms in the field of Healthcare, where I believe AI has made the most impact. My present and upcoming projects are all related to this field, and the world will see amazing powerful cognitive technologies through which we give artificial communication abilities to those who cannot communicate naturally and so much more!
You clearly have an ability to understand and convey quite complex concepts - but what is something that you find particularly difficult to do?
I find it difficult to take concepts that are already conceptually difficult, like AI, but and at the same time also have unnecessary fear around them, to simplify and distill them down so that not only can people understand the concepts, but also why they shouldn’t be afraid of those new concepts. See, AI already has a lot of fear in terms of overpowering humanity, replacing our jobs, etc., but it’s also a difficult topic to explain – you can’t easily describe it to the average Joe by talking about the backend technologies! You’ve got to distill it down.
What are your passions & hobbies outside of Machine Learning & AI?
I love to travel, visit new places and meet new people and make friends. I’m very passionate about teaching, I love to teach, therefore wherever I’m invited for a keynote, I make sure to do workshops as well. I also love to go for long runs, in fact my newly found interest is cricket, I think I’m going to spend a lot of time learning and playing! You’ll get to hear more about this very soon!
We hope you enjoyed the interview - please share using the social buttons below and look out for our next edition of SPOTLIGHT!