Introduction
Quantum computing and quantum machine learning – most of us have come across these concepts at some point without getting the opportunity to delve deeper. But what if I told you that these could potentially disrupt the way we see and use technology?
We are joined by Dr. Mandaar Pande in episode #14 of the DataHack Radio podcast, where he navigates us through the wonderfully complex world of quantum computing. Here’s a mind-blowing fact to give you a taste of what to expect:
“The number of bits in a 300 qubit quantum computer will be more than the known atoms in the universe.” – Dr. Mandaar Pande
Before I personally met Dr. Mandaar at DataHack Summit 2018 (where he also spoke on this subject), I only had a vague sense of what quantum computers are and the gigantic amount of power they can process. But as you’ll soon find out, there’s a lot more that goes on behind the scenes that one might never have thought of.
I have briefly covered the main topics discussed in this episode but the true joy and knowledge lies in listening to Dr. Mandaar himself.
Subscribe to DataHack Radio today and listen to this, as well as all previous episodes, on any of the below platforms:
Dr. Mandaar Pande’s Background
Dr. Mandaar holds a Ph.D degree in theoretical physics from the University of Hyderabad, with a specialization in non-linear optics. After completing his Ph.D in 1994, he took up a post as a lecturer at BITS Pilani for the next four years in the EEE department (electrical engineering).
Experienced folks will recall that it was in the late ’90s when IT started picking up steam in India, and Dr. Mandaar decided to take the plunge and explore other avenues outside of academia. He joined Tech Mahindra in 1998 as part of the modeling and simulation centre. There, he worked in the capacity of Group Head and Principal Consultant in the area of Performance Engineering and Management.
Following a 12 year stint at Tech Mahindra, he spent 7 years at Wipro – first as a Lead Architect, then as the Global Practice Head for Performance Engineering – Quality Engineering and Testing. Dr. Mandaar’s experience in this field, as you can tell, is incredibly rich and not many folks come close to rivalling his experience and know-how.
But for Dr. Mandaar, it felt inevitable that he would return to academia at some point and he joined Symbiosis as the professor of IT last year.
Interest in Quantum Computing and Quantum Machine Learning
So where does his interest and passion for quantum computing and machine learning fit into the picture? Well, towards his final years with Wipro, he got some exposure to the digital way of working (and data science was a big part of that). He built up an interest while working there, and kept that up during his transition back to academia with Symbiosis.
His Ph.D in quantum optics obviously helped while he pursued quantum computing. But what is this field exactly? And how does it tie into machine learning? Let’s hear that from Dr. Mandaar himself:
“Quantum Computing is a field that is at the intersection of quantum physics, information science, as well as function theory. And one of the largest applications of quantum computing in the near future is going to be quantum machine learning. “
What are the key characteristics one needs to have in order to learn Quantum Computing?
This is a tough one, and a question I have been wondering about ever since I heard about this subject. Below are the two key points Dr. Mandaar mentioned:
- A solid foundation (and interest) in mathematics and physics. Dr. Mandaar’s Ph.D in theoretical physics was a major boon in that regard
- The ability to think differently and show patience to stay with a challenge for a long time. People tend to give up after a certain point of time and that just won’t do in a field as nascent and research intensive as this
There’s a lot more to it, but if you don’t possess a solid base in these two aspects, it’s going to be next to impossible to make headway.
But what is Quantum Computing in the first place?
The current version of digital devices we use, like computers and smartphones, use metal chips in them (which are based on integrated circuits that are in turn based on transistors). Whatever computing we do today, including the data we capture and the analysis we perform, is done using the 2 bits we see in these transistors – 0 and 1. Any algorithm that we write eventually gets broken down into 0 and 1 at the machine level.
At the most fundamental level, the physical principles that govern the nature in quantum computing follow the laws of quantum mechanics. So then what is quantum mechanics? It’s a theory in physics that describes nature at the smallest level (atoms and molecules). At this really basic level, these particles behave very differently.
Now, there are a couple of things that are very important to understand:
- Unlike classical computers which are based on bits, quantum computers are based on qubits (quantum bits)
- Any instruction we pass to a classical computer will be solved sequentially (despite the rise of GPUs and TPUs). In quantum computing, the concept of parallelism comes into play
This is just a taste of what Dr. Mandaar described in the podcast. He broke down this complex topic into easy-to-digest bits of information using examples. If you’ve ever wondered how quantum computers work from the ground up, this section will feel like you’ve hit the jackpot.
Can we put a Timeline on when Quantum Computing will become Mainstream?
The general perception is that it might take up to 15-20 years before we see quantum computing becoming mainstream and getting democratized.
But as Dr. Mandaar pointed out, these things are not written in stone. A big breakthrough has the potential to throw the doors wide open and accelerate the timeline (as we have seen with deep learning and AI in recent times). But until that happens, quantum computing will remain the realm of the big organisations like IBM.
Impact of Quantum Computing on Machine Learning
According to Dr. Mandaar, most of the classic machine learning and deep learning algorithms have a quantum equivalent now. He took the example of a perceptron, which is also available as a quantum perceptron. Further, Dr. Mandaar mentioned four ways how one can look at Quantum ML:
- CC: Running a classical ML algorithm on a classical computer (this is where we are currently)
- CQ: Running a classical ML algorithm on a quantum computer
- QC: Running a quantum ML algorithm on a classical computer (this is being researched nowadays)
- QQ: Running a quantum ML algorithm on a quantum computer (this is still a long way off)
End Notes
That was a power-packed episode! There was a lot Dr. Mandaar told us about a subject most of us have vaguely heard of in our professional careers. Apart from the points I’ve mentioned here, the conversation also covered topics like challenges in quantum computing, the areas where classical computers could potentially be better than quantum computers, a very specific list of requirements one needs to learn in order to gain traction in this field, what a quantum internet is (really cool concept!), among other things.
Senior Editor at Analytics Vidhya.
Data visualization practitioner who loves reading and delving deeper into the data science and machine learning arts. Always looking for new ways to improve processes using ML and AI.
