Briefly about quantum algorithms and their families
One of the goals of the development of quantum technologies is to accelerate
A starting point in the development of quantum computingconsider the 90s. Even then, the first algorithms for quantum computers existed, for example, the Deutsch algorithm, but they were primarily of academic interest. Somewhat later, American scientists Peter Shor and Lov Grover developed two essentially different, but very valuable quantum algorithms from a practical point of view. Shor's algorithm convinced the world that quantum technologies have very real applications and that the quantum computer threatens modern methods of cryptographic information protection. And Grover's idea indicated that with the help of quanta, you can search through unstructured data faster than with the help of classical technologies (and, as it was later proved, with the maximum acceleration available in this task).
Both discoveries have not yet brought direct practicaleconomic results. Nevertheless, the era of studying quants as an applied tool began with them - and other algorithms that are more applicable in reality are based on them.
For example, the Zalka-Wiesner algorithm will helpcreate new drugs and new materials. Ambainis algorithm - analyze data, such as images and texts. Another of the most notable inventions is the Harrow-Hassidim-Lloyd algorithm, which solves a system of linear equations on quantum computers. This will greatly advance the processing of big data, the modeling of complex systems, and ultimately may be another step towards the creation of strong artificial intelligence.
New ideas for quantum algorithms emergeregularly: most are listed on the Quantum Algorithm Zoo website, and implementation examples can be peeped on Github. In addition to the United States and China, Spain is a big contributor to this trend — it leads the development of financial algorithms thanks to the active movement of the startup Multiverse Computing. England has successful solutions for speech recognition and chemical process modeling. Soon, India will also make a name for itself — numerous university courses in quantum areas are being launched there.
What still slows down quantum progress
So, a quantum algorithm.Let's start with the more understandable word “algorithm,” which means a certain sequence of actions, a recipe for obtaining a certain result. The word "quantum" in this case adds to our ability to have certain logical operations that are not available in the classical world. Thanks to the phenomena of quantum physics - superposition and entanglement - using logical operations on the states of quantum objects, it is possible to create complex quantum states that would classically require a lot of numbers to describe. For example, to describe the state of 50 qubits, in general, up to 2^50 numbers are required. If you correctly use this property of quantum systems and calculate the sequence of logical transformations, and then implement them on a quantum computer, you can find the answer to the desired question in the case where a classical computer would require a lot of time and memory resources.
It would seem that create yourself different algorithms fordifferent tasks, and in theory there are already dozens, if not hundreds, of them. But when it comes to practice and significant (i.e., exponential) acceleration, the development of quantum algorithms runs into certain barriers. And therefore, the emergence of new quantum algorithms that can lead to radical changes is rather rare: every decade there are only 5–10 new interesting quantum algorithms.
The main difficulty is related to the fact that in the minds of mathematicians and theorists, algorithms run on an ideal computer. However, real quantum computers are not at all error-proof.
The obstacles boil down to the fact that it is very difficult to create a powerful error-free quantum system:
- Quantum objects are very sensitive to the slightestenvironmental changes: even minimal leakage of quantum information into the environment during the execution of the algorithm can lead to distortion of the final answer. And this despite the fact that the conditions for the stable operation of qubits in certain cases must be maintained very specific: for example, the temperature must be close to absolute zero, that is, correspond to -273.15 ° C.
- Noise affects the calculation process - due toconstant exposure to errors, the number of operations that can be implemented in a quantum algorithm is now limited to several dozen, while thousands of operations are required to win when solving industrial problems.
- So far, it has not been possible to create the required numbererror-free, that is, logical qubits. For example, to calculate the risks of a company operating in the securities market, a minimum of 200 logical qubits is required, to crack cryptography - about 6 thousand, and to predict the value of financial derivatives in real time - 7.5 thousand. For even more complex tasks like fast hacking of the network, bitcoin or modeling of new materials requires from several tens to hundreds of millions. Whereas today, quantum devices do not have more than units of error-proof qubits.
To deal with quantum noise, scientistsoffer different correction methods - and they really help reduce the level of errors in the system. But their effect on calculations will be significantly reduced only after a certain time. Forecasts vary: IonQ puts it at 2025, and Google counting on 2029.
How soon quantum algorithms will change our lives
Quantum computing hasn't hit the mainstream yet, butare already affecting living standards—or will soon be. For example, the D-Wave quantum computer has shown the ability to significantly speed up the work of the Los Angeles seaport (while research is being conducted on test data), Daimler and IBM Quantum are about to create quantum batteries for electric vehicles, and RKC and the QBoard project have developed a quantum algorithm for safe disposal of nuclear waste in the interests of Rosatom.
Most quickly - within 2-3 years - they will manifest themselvesvariational algorithms for solving optimization problems that are widely applicable in almost any industry. This is a hybrid computing class: the quantum part is responsible for changes in state, and the classical part helps stabilize this process. And, since hybrid models are much simpler to implement than fully quantum ones, the power of existing quantum computers is sufficient for them. RKC and the QBoard project use quantum algorithms to solve chemical modeling problems for the benefit of the automotive industry. However, the development of variational algorithms faces a whole set of challenges. On the way to their implementation, the entire field of quantum computing is in for interesting discoveries.
Read more:
Oldest Cyrillic text ever found
"Webb" found traces of incredibly huge stars: they died at the dawn of the universe
Strange sounds recorded in the Earth's stratosphere: how scientists are trying to understand their nature
Cover photo: Image by starline on Freepik