Introduction to Quantum Computing — For Newbies and IT Professionals — Part 1
Enter the world of Quantum Computing with a 3-part Introduction
Quantum computing is poised to change the computing landscape, unleashing solutions for a plethora of super complex, hitherto unsolvable or time-demanding problems. Google’s demonstration of quantum supremacy last year was just one of the steps in this direction. There are multiple startups, multi-nationals and government led projects which are already making major strides towards a quantum-powered world.
As more research gets done in this area, it’s only a matter of time but yes, it will take some time, until we see use of quantum computing for desired applications. This series of articles aims to introduce you to the quantum world, explain why there is a lot of excitement in this space, what are the different ways to build a quantum computer, what are its limits, who are the current big players, how you can get started with quantum programming etc.
Let’s start with quantum supremacy.
What is quantum supremacy?
Quantum supremacy is a term to represent the idea/fact that the quantum computers of the future will be able to perform certain tasks exponentially faster than the conventional classical computers. It is capable of solving a lot of problems currenlty intractable on classical computers / supercomputers. Google demonstrated quantum supremacy last year with their purpose-built “machine”, equipped with a 53–qubit (more on qubits later!) Sycamore processor. Note that I call it a machine here, just to point out that a quantum computer won’t look anything at all like a classical computer (at least not in the near future!), given the controlled environment it needs to operate in.
In the quantum supremacy experiment done by Google, they exhibited how a certain task (specifically, sampling from outputs from “random circuits”) can be performed by a quantum computer much faster (3 minutes) compared to a supercomputer, specifically the Summit supercomputer, which will take between 3 days and 10,000 years! If you are wondering why there is such a discrepancy between the time it will take, you can read this article where Google’s claim of 10,000 years is disputed by IBM, who claim the task can be done in 2.5 days on a classical system.
You might ask: Can’t we simulate quantum computing on classical computers? Yes we can and it has been done for a lot of years. Infact, that has been one of the primary use-cases for supercomputers. However, as we try to simulate larger quantum systems with more qubits, costs and resources required grow exponentially. For example, you need 16 PB of memory to simulate a 50-qubit quantum system in order to cover all the states. Thus it slowly becomes infeasible to be able to simulate quantum computers on classical computers.
Hindrances to mainstream Quantum computing
So now that we have proven quantum supremacy, why are we still using classical computers? Well, quantum computing is still in its infancy. There are still a lot of existing problems to overcome till it gets more mainstream. Some of these problems are:
- Quantum decoherence — You can understand quantum decoherence as a loss of information in qubits due to their interaction with the environment. Quantum coherence (opposite of quantum decoherence!) has been shown to be equivalent to quantum entanglement, which is one of the main properties of quantum systems that quantum computing exploits/uses.
In order to keep quantum coherence, quantum computing has to be done in a very controlled environment, i.e. extremely low temperature, vacuum-like, void of external electro-magnetic waves etc. It is however very hard to properly isolate a quantum system, which leads to loss of quantum coherence. During quantum decoherence, quantum superposition deteriorates causing the loss of information, in which case there is no speed up compared to classical computing.
Quantum decoherence is one of the main reasons why it has been so hard to build quantum computers using Gate-based models. More on quantum superposition, entanglement, gate-based and other models later!
- Error-correction required in the qubits — This is actually a consequence of the quantum decoherence phenomenon. The qubits are very noisy and hard to control. This is why the current era is know as the Noisy Intermediate-Scale Quantum (NISQ) era.
In order to perform error correction, an error rate is used to model how “good” a qubit is. The error rate represents the probability of error (quantum bit/phase flip) during execution of a gate/operation. Further error correction is performed by using redundant encoding (more qubits used to save the same information as one qubit) and entanglements to fight the effect of noise. However using error correction means that you use a lot of qubits not for the actual work, causing a qubit overhead. As you add more and more qubits into the system, the error rate goes up. If you go above the error correction threshold, it’s not possible to do error correction anymore.
- There is no established way to feed a lot of data to quantum computers. Quantum computers, as of now, don’t handle big data. However where quantum shines is when the number of inputs are small but the computations explode as you start examining relationships or dependencies in the data.
So it looks like there is still time till Quantum computing becomes mainstream. Let’s however look at some of the exciting applications of quantum computing .
Applications of Quantum computing
There have been quantum algorithms like Shor’s algorithm (integer factorization), Grover’s algorithm (finds with high probability the input for a particular output value / quantum search) etc., already developed which can be used on quantum computers solving problems in cryptography (Shor’s algorithm), improve search performance using Grover’s algorithm, providing quadratic/exponential gains in search performance, machine learning, large-scale weather and climate forecasting, optimization, improved drug design and personalized medicine in pharmaceuticals, wave-based geophysical processing to locate new drilling sites for oil and natural gas, among many other applications.
Let’s take an example of just the finance industry and see what applications are possible in the quantum computing world.
- Fraud detection — In the quantum era, you can better detect outliers even in the face of a growing number of variables, thus improving the accuracy of your models.
- Derivative pricing — Improvement in the performance of complex options pricing.
- Portfolio optimization — Make better assumptions and optimizing using a lot more diverse assets than ever before.
- Other use-cases include better capital allocation and risk assessment , improved asset liability management, find more profitable trades, further machine learning/optimization use-cases etc.
In this article, I explained quantum supremacy over classical computing and what are the current hindrances towards it becoming mainstream. We also looked at some of its possible applications in different domains. In the next two articles, we will go through the quantum concepts like qubits, superposition, entanglement, teleportation, quantum annealing, different models of building quantum computers, development frameworks etc. You can find the next article here. Hope to see you there!