For many of us, artificial intelligence as a concept seems very new and recent, especially when it comes to India, even though the technology had its origin over half a century ago. In that sense, one has to say, Nikhil Malhotra, who currently heads Global Innovation and Makers Lab at TechMahindra, began his AI journey at a time when AI wasn’t “even a thing” in this country. 

Starting at IBM back in 2006, with a focus on computational analysis, Nikhil has led and transformed research and development across various organisations, especially after moving to TechMahindra in 2007. 

“In 2014 I was in the US, and I had my chance when our CEO asked us to start an R&D division within the organisation, where we can have the research and analysis, as well as the larger perspective on the AI,” recollects Nikhil. 

“2014 also happened to be a critical year as everything came down right. India was growing as an economy; AI systems were proving its effects. That is the journey I started in India, and that is the journey I am continuing,” added Nikhil. 

The initial phase of AI research at TechMahindra Maker’s Lab began with a focus on supervised and unsupervised reinforcement learning exploring new techniques and algorithms. However, in recent time Nikhil and his team are also researching facets of neuroscience. 

“The key question we are looking at is how can we utilise a smaller amount of data by looking at how our brain is working. For example, can we look into those neural pathways and turn it into an algorithm as well?”

The future of AI is in quantum computing

Apart from neurosciences, another emerging technology that could transform AI is quantum computing. Despite being a technology in its nascent stage, Nikhil and his team have been working towards unleashing its superpowers.

“There is a lot of development happening with quantum computing. Quantum computing would help us through three principles. And these three principles are working at a subatomic level and not at a classical level,” Nikhil explained. 

“In fact, the world of subatomic particles is very different, that Richard Feynman once said that if you are not confused about quantum computing or quantum physics, then you are not grasping anything. That world is complex.”

“There is a historical context to this,” explained Nikhil, “the origin of quantum computing goes way back to the days when it was finally discovered that light could behave as both wave and particle.”

It was in 1801; physicist Thomas Young conducted the now-famous “double-slit experiment” in which he shot a beam of light toward a barrier with two slits in it. Instead of forming two lines on the screen behind as expected due to the particle nature of light, the beam formed a pattern of interference as if a wave had been pushed through the two slits.

A century later Geoffrey Ingram Taylor was able to repeat the experiment using a single photon - a single light particle. The wave interference pattern still appeared in his tests as well, suggesting the duality of light or light particle’s characteristic to behave as both wave and particle. This has resulted in the establishment of one of the most fundamental concepts behind quantum physics, the wave-particle duality. 

The first of three fundamental principles of quantum computing is superposition which states that if a physical system may be in one of many configurations of particles or fields, then the most general state is a combination of all of these possibilities, where the amount in each configuration is specified by a complex number. 

“In a classical physics worldview, we know that computing has been based on zero or one, like the light switch which can always be turned on and off. However, in the quantum world, we look at the subatomic particles, which can exist as waves as well as particles at the same time. As a result, there can be more zeroes and more ones at the same time. It gives us immense computing power,” he explained. 

“If I have two picks, then I can have one of the four combinations in computing, namely 00, 01,11, and 10. However, in the quantum level or at the subatomic particle level, you can have all four at the same time and construct solutions for the problem we give.”

Another critical characteristic of quantum computing is entanglement. “Entanglement seems to be the spookiest principle of all times, and that is what gets people confused,” according to Nikhil. 

Quantum entanglement is a quantum mechanical phenomenon in which the quantum states of two or more objects have to be described with reference to each other, even though the individual objects may be spatially separated.

“Let’s say somewhere two-particle gets created as entangled. I take one of those particles and keep it in my pocket, and I take the other particle and take it to the end of the universe. Now if I measure the particle in my pocket, then the particle that I kept at the end of the universe will also change based on that measurement. Now, what’s fascinating here is that this defies our current view of space and time, and this same principle is being utilised in quantum computing.”

The third principle of quantum computing is quantum tunnelling, which occurs when particles move through a barrier that, according to the theories of classical physics, should be impossible to move through. The barrier may be a physically impassable medium, such as an insulator or a vacuum, or a region of high potential energy.

“Let’s take a ball as an example,” Nikhil explained, “when you are throwing the ball over the hill, depending on the kinetic energy the ball has, it will either move, or it will stop.” 

“Similarly, when throwing a ball at a wall, it bounces back. However, in a quantum world, because of the wave nature, a small percentage of particles can come on the other side. The ball here in the quantum world can cross the barrier of the wall and the hill, and they can come out at the other side.”

By using tunnelling, computation problems get solved faster and give rise to quantum machine learning. The key for quantum computing to bring the effect it promises on paper is its integration with AI.

While machine learning algorithms are used to compute immense quantities of data, quantum machine learning, a product of integration between AI and quantum computing, increases such capabilities intelligently, by creating opportunities to conduct analysis on quantum states and systems.

“Why I am excited about quantum comporting is that it is no longer science fiction. There are practical sides to it, and real use cases are coming up.”

Coming to India, this year, we saw a stream of new initiatives announced such as the National Quantum Computing Mission by the government. However, organisations like TechMahindra has been pioneering this technology for many years for now in the private sector.

“A lot of works has been done in India. There are many research institutions. From TechMahindra’s perspective, we have taken it as one of the biggest bets. Because some years back, we felt that a research centre with AI should also broker with quantum, quantum-AI enablement will be the next big leap forward.”

Fighting COVID with AI 

Ever since the beginning of this year, humanity has been facing a long drawn battle against the COVID19 pandemic. One of the major highlights of this crisis has been how much we started to resort to automation and AI tools. From diagnostic, crisis management, to drug discovery to reopen it of economic activities, AI has been playing a crucial role. 

“AI has been playing a huge role, I believe. If anything can defeat this virus at this point in time, then it has to be the combination of two factors. First, a combination of computational analysis, and secondly getting these result to the hands of doctors and researchers.”

Throughout the world scientist and research have been dedicating a vast amount of time and resources to solve the mystery of the COVID19 virus, and crucial to unlocking its mystery lies with its protein folding structure. 

“I think what we initially figured out with the research, and we are continuing to do that is the research of protein analysis, the research on protein sequences, figuring out what molecular compounds can act. Importantly now everybody knows that the virus connects to the enzymes of the human body using a particular spike, and that spike can be nullified,” stated Nikhil. 

“Now, how do we nullify that? It is through going back to the basics and trying it out through trial and error. And that will take a medical system much longer in terms of time, and computational analysis and AI will definitely help,” he added.

Furthermore, there has been an increased interest in designing preventive measures using AI. 

“For example, proximity or getting a lung scan or figuring out who has that virus and who is liable to affect it are the key preventive solutions AI can develep.. Another challenge for the organisations as well as the government is to segregate out what clusters in the city, district or state, would grow based on the virus. Analysis based on other factors such as the population density, people travelling in and out from a specific cluster, people who can be a possible carrier of the virus, how much they can spread, are also critical.”

“Another facet that we are looking for is more predictive in nature. We are looking at how AI can predict based on genomic DNA sequences, as to what forms of virus virulence can occur in the near future.”

“Is there a way that the virus would become virulent after it goes down? Everybody is talking about that may be temperature can pin the virus down. But that has been proved wrong, and we now ask questions like is it going to be more virulent, or is it going to be less in its effect. I think computational analysis can really help us here, especially, with AI is an effective tool for this.”

AI in the post-COVID world

"Today, even though we are nowhere close to seeing the end of this crisis, there has been a lot of debate on the nature of much-changed the post-COVD world. According to Nikhil Malhotra, there will be a huge transformation. “Once we get over this phase of isolation, people will realise that we can do a lot more with AI within our ecosystems."

“One of the things that people lack, or organisations typically lack these days with regard to AI is the trust factor. We see a lot of people talking about AI, but we see many people saying why they cannot trust machines taking decisions. That trust value factor coupled with how AI can become more ethical is what will bring the biggest economic impact in India. Industries that will change in India post this COVID will be health care, education, telecom and network influenced by AI. That is going to be the biggest economic impact,” he expressed his optimism.