Cracking the uncertainty around quantum computing

Today’s leaders are inundated with the disruptive power of quantum computing and its potential applications in AI, machine learning and data science. Gartner data reveals that by 2023, 95% of organisations researching it will utilise quantum-computing-as-a-service (QCaaS) to minimize risk and contain costs. Also, 20% of organisations will be seen budgeting for quantum computing projects, compared to less than 1% today.

We, Aravind Ajad Yarra, fellow, Wipro Limited and Saji Thoppil, fellow and chief technologist – cloud and infrastructure Services, Wipro Limited, bring you the basics of quantum computing and demystify some of its unknown facets in today’s evolving scenario.

Let’s look at the commonly asked questions:

Q: Can you define what quantum computing is in simple words?

A: Most of us would have read quantum mechanics at high-school level physics and probably been baffled by its strange characteristics. Quantum mechanics is the physics that applies at atomic and subatomic levels. Thought of using the physics of quantum mechanics to computing is what has led to quantum computing.

Our present-day computing is largely based on Boolean logic, represented using binary bits, which assume the value of either 0 or 1. Quantum computing, on the other hand, uses quantum bits (qubits), which behave differently from classic bits and use quantum superposition state where each qubit can assume both 0 and 1 at the same time.

To get better clarity, I suggest reading this short article on quantum computing.

Q: Why is quantum computing the new buzzword in business today?

A: Quantum computing is one of the most exciting developments in recent computing history. For years, Moore’s law has been helping us to keep the innovation cycle in computing going and push the boundaries of what computing can offer to business, so much so that software is what is driving digital businesses. With Moore’s law reaching its saturation point, everyone is eagerly looking for what’s next in computing. This is seen as something that can keep the computing innovation cycle going, hence this buzz.

If you hear the general hype, you might believe quantum computing might replace classic computing soon. However, that is far from reality. The superposition property that we mentioned earlier gives quantum computing some unique capability that traditional computing doesn’t have. Simply put, qubit superposition allows quantum computing to solve certain classes of problems promptly, which might otherwise take years for classical computers.

IBM achieves highest quantum volume to date

IBM has established a roadmap for reaching quantum advantage and concluded that: for significant improvement over classical systems, the power of quantum computers must double every year. Read here

Q: What problems can it address that existing technologies are not able to solve?

A: Quantum computers are not bigger or faster versions of existing computers. Quantum computing is fundamentally different from existing computing. The problems for which quantum computers are most useful are problems that classical computers are not good at.

Some of the classes of problems that quantum computers currently look at are optimisation problems, for example, addressing the classic travelling salesman problem. As the number of cities that have this problem increases, classic computers find it exponentially hard to find an optimum solution. Quantum computers proved very useful for these classes of problems. Solving such problems make quantum computers super useful in areas like gene analysis, drug discovery, chemical synthesis, weather simulations, newer types of encryption, unstructured search, and better deep neural networks, to name a few.

What is AI? A simple guide to machine learning, neural networks, deep learning and random forests!

What is AI? Information Age has created a simple guide to AI, machine learning, neural networks, deep learning and random forests. Read here

Q: What approaches are currently available?

A: There are two major approaches to quantum computing that are currently in use: circuit-based computers (aka universal quantum computers), and adiabatic computers.

Universal quantum computers are based on logical gates and work similar to the underlying logic foundations of classical computers. Hence, universal quantum computers are extremely useful for computing problems improving on our current knowledge base of solutions. However, qubits required for universal quantum computers are extremely difficult to realise physically because qubit instability makes it hard to produce universal quantum computers.

Adiabatic computers are analog, but are easier to produce. These are more relaxed with respect to qubit state stability. Hence, it is easier to produce 1000s of qubits on adiabatic computers. However, adiabatic computers can be used for limited use cases such as optimisation problems.

Q: Which model should an enterprise bet on?

A: While most platform companies that are working to build quantum computers are taking bets on one or the other, enterprises can probably explore both of the models. While adiabatic computing is limited, there are production-ready adiabatic computers using real quantum bits (such as those from DWave), as well as digital annealers, which use digital qubits (from Atos and Fujitsu).

Are emerging technologies set to transform business?

It’s emerging technologies month on Information Age, that means augmented and virtual reality, quantum computing and blockchain. Read here

Circuit-based quantum computers are much more general purpose. While these have more utility for enterprises, no production-grade problems can be currently solved with the current state of these machines. I would suggest exploring both classes of computers, based on the case that one is trying to solve.

Q: How to identify the use cases for quantum computing?

A: The best way to start with identification of use cases for quantum computing is to explore areas where classic computers are currently not good at. Optimisation problems are the best starting point for most enterprises. Based on the industry, different kinds of optimisation use cases can be considered for exploring quantum computers. These could be risk modelling, inventory or asset optimisation, among others.

Cryptography is another area where robust use cases can be identified by enterprises. Quantum computers, when production-ready, can potentially break current methods of encryption, leading to exposure of sensitive data. Identifying data that is very sensitive and has longer term value, and considering safe encryption methods using quantum key generation and distribution are other ways in which it can be used.

Machine learning is also a very promising use case. Quantum machine learning, as it is called, can use special purpose quantum circuits that can significantly boost the efficiency of machine learning algorithms.

Q: Which industries are the early adopters of quantum computing?

A: Industries that are process-centric, such as pharmaceuticals and oil & gas exploration, are the early adopters. These industries can benefit from quantum computing in complex optimisation problems they need solve from time to time.

Apart from these asset-heavy industries, the manufacturing industry is also actively exploring quantum computing. Banks and other financial services companies, which have risk modelling needs, also rely a lot on quantum computing.

Q: Any examples of real-world scenarios where quantum computing has made an impact?

A: It is probably too early to talk about real-world scenarios where quantum computers have made an impact. While there are demonstrations by research labs to use quantum communication methods to send instant data transfer from satellite and breaking various encryption methods, these still look good in labs.

The reason for this is the current state of reliability in quantum computers. Qubits are highly sensitive, and they are prone to errors. Error correction methods that we currently use reduce the effective working qubits, but early results have been seen with digital annealers, which simulate adiabatic quantum computing using traditional digital computers.

Wipro’s Topcoder, for example, is currently working with Fujitsu to run crowdsourced challenging using Fujitsu’s digital annealer to solve real-world problems. Additionally, Airbus has been running open innovation challenges to solve some of its problems using quantum computing.

Q: Other than quantum computing, what are the other quantum technologies that the enterprise should keep an eye on?

Quantum technologies also has appeal in the areas of communication, cryptography, sensors and measurements. Unlike quantum computing, where practical use cases are still in exploratory stages, these areas have industry-ready products that enterprises can put to use.

Quantum communication takes advantage of the nature of photons in flight and is able to detect if a photon has reached the recipient uninterrupted; this can ensure secure communications.

While quantum key generation (QKG) is used to generate truly random keys, quantum key distribution (QKD) is used for securely distributing keys. Both of these are essential for using a one-time pad cryptography technique, which is considered the holy grail in encryption.

Random numbers for the quantum computing world, has Crypta Labs really cracked it?

Generating true random numbers for the quantum computing era, or indeed the pre-quantum era, is the aim. Crypta Labs reckon they have cracked it. Read here

Additionally, quantum sensors have niche applications where there is a need for highly accurate measurements of gravity, electric fields, time, position and magnetic field. In a fiercely competitive world, we can expect more enterprises wanting to leverage these to create unique offerings.

Q: How do you suggest one should build the business case for quantum computing in an enterprise?

Given the nature of its evolution, it is hard to make an upfront business case for quantum computing. However, given the potential, I suggest that the business case be made in two parts.

The first part is to focus on near-term (1-2 years) use cases such as optimisation and encryption by using digital annealers for optimisation and photon-based ASICS for key generation. Digital annealers, or even simulators running on cloud, can solve several practical optimisation problems.

On the other hand, centres of excellence can be set up, leading to building expertise and solving relevant problems. Returns from these investments would set the stage for the second part, focusing on mid & longer term (2+ years) use cases, such as exploring machine learning and unstructured data search as part of centres of innovation and open innovation communities with small investments, but with longer period on returns.

Written by Aravind Ajad Yarra, fellow at Wipro Limited, and Saji Thoppil, fellow and chief technologist – cloud and infrastructure Services at Wipro Limited