What aspects of quantum computers, if any, can help to further develop Artificial Intelligence?


Quantum computers are super awesome at matrix multiplication, with some limitations. Quantum superposition allows each bit to be in a lot more states than just zero or one, and quantum gates can fiddle those bits in many different ways. Because of that, a quantum computer can process a lot of information at once for certain applications.

One of those applications is the Fourier transform, which is useful in a lot of problems, like signal analysis and array processing. There's also Grover's quantum search algorithm, which finds the single value for which a given function returns something different. If an AI problem can be expressed in a mathematical form amenable to quantum computing, it can receive great speedups. Sufficient speedups could transform an AI idea from "theoretically interesting but insanely slow" to "quite practical once we get a good handle on quantum computing."

  • 1
    $\begingroup$ Adding to this answer, matrix multiplication is the backbone of most Machine Learning applications today. Anything that uses a GPU today could tomorrow use a quantum computer bringing us that much closer to AI. $\endgroup$
    – Harsh
    Aug 3 '16 at 1:03
  • 2
    $\begingroup$ ML is subset if AI. Deep ML is subset if ML. Hence we are not building any AI by multiplying matrices. We simply cut the best pieces out of a corpse like piranhas. For publications scores degrees grants tenures and money. I really doubt an actual AI is going to have anything to do with speed of multiplying matrices. Human brain produces intelligence effortlessly using infinitely slower apparatus compared to a quantum computer. $\endgroup$ May 28 '17 at 22:00

Until we can make a quantum computer with a lot more qubits, the potential to further develop AI will remain just that.

D-Wave (which has just made a 2,000+ qubit system around 2015) is an adiabatic quantum computer, not a general-purpose quantum computer. It is restricted to certain optimization problems (at which its effectiveness has reportedly been doubted by one of the originators of the theory on which it is based).

Suppose that we could build a 32 qubit general-purpose quantum computer (twice as big as current models, as far as I'm aware). This would still mean that only 232 possibilities exist in superposition. This is a space small enough to be explored exhaustively for many problems. Hence, there are perhaps not so many problems for which any of the known quantum algorithms (e.g. Shor, Grover) would be useful for that number of bits.

  • $\begingroup$ "D-Wave (which has just made a 2,000+ qubit system around 2015)" This statement is misleading at best. Be aware that D-Wave has claimed to create a computer using adiabetic quantum annealing. This computation model is significantly different than other quantum computing models. For example, I'm not aware whether Shor and Grover work on this model! So, talking about "2,000+ qubits" is a bit misleading: computers in the model where we care about the qubit count have something around 50 qubits as the current frontier. $\endgroup$ Mar 21 '18 at 13:28
  • $\begingroup$ Also note that there are experts that do not believe adiabetic quantum annealing can give significant improvements on the classical computing technique of simulated annealing. $\endgroup$ Mar 21 '18 at 13:29

Quantum computers can help further develop A.I. algorithms and solve the problems to the extent of our creativity and ability to define the problem. For example breaking cryptography can take seconds, where it can takes thousands of years for standard computers. The same with artificial intelligence, it can predict all the combinations for the given problem defined by algorithm. This is due to superposition of multiple states of quantum bits.

Currently, quantum computers are still in the early stages of development and can perform complex calculation. There are already technologies like D-Wave systems which are used by Google and NASA for complex data analysis, using Multi-Qubit type quantum computers for solving NSE fluid dynamics problems of interest or global surveillance for military purposes, and many more which we're not aware.

Currently there are only a few quantum computers available to the public, like IBM Quantum Experience (the world’s first quantum computing platform delivered via the IBM Cloud), but it's programming on quantum logic gates levels, so we're many years behind creating artificial intelligence available to public. There are some quantum computing languages such as QCL, Q or Quipper, but I'm not aware any libraries which can provide artificial intelligence frameworks. It doesn't mean it's not there, and I'm sure huge companies and governments organisations are using it for their agenda to outcome the competition (like financial market analysis, etc.).


Direct Answer to Your Question:--

The field where quantum computing and A.I. intersect is called quantum machine learning.

  1. A.I. is a developing field, with some background (ala McCarthy of LISP fame).

  2. Quantum computing is a virgin field that is largely unexplored.

A particular type of complexity interacts with another type of complexity to create a very rich field.

Now combine (1) and (2), and you end up with even more uncertainty; the technical details shall be explored in this answer.

Google Explains Quantum Computing in One Simple Video: Google and NASA's Quantum Artificial Intelligence Lab


IBM is an authority:--

IBM: Quantum Computers Could Be Useful, But We Don't Know Exactly How

Quantum machine learning is an interesting phenomenon. This field studies the intersection between quantum computing and machine learning.


"While machine learning algorithms are used to compute immense quantities of data, quantum machine learning increases such capabilities intelligently, by creating opportunities to conduct analysis on quantum states and systems." Wikipedia contributors. — "Quantum machine learning." Wikipedia, The Free Encyclopedia. Wikipedia, The Free Encyclopedia, 7 Oct. 2019. Web. 11 Oct. 2019.

Technical Mirror:--

This particular section on the implementations is worth noting:--


" ... This dependence on data is a powerful training tool. But it comes with potential pitfalls. If machines are trained to find and exploit patterns in data then, in certain instances, they only perpetuate the race, gender or class prejudices specific to current human intelligence.

But the data-processing facility inherent to machine learning also has the potential to generate applications that can improve human lives. 'Intelligent' machines could help scientists to more efficiently detect cancer or better understand mental health.

Most of the progress in machine learning so far has been classical: the techniques that machines use to learn follow the laws of classical physics. The data they learn from has a classical form. The machines on which the algorithms run are also classical.

We work in the emerging field of quantum machine learning, which is exploring whether the branch of physics called quantum mechanics might improve machine learning. Quantum mechanics is different to classical physics on a fundamental level: it deals in probabilities and makes a principle out of uncertainty. Quantum mechanics also expands physics to include interesting phenomena which cannot be explained using classical intuition. ... " — "Explainer: What Is Quantum Machine Learning And How Can It Help Us?". Techxplore.Com, 2019, https://techxplore.com/news/2019-04-quantum-machine.html.

Business Applications and Practical Uses:--

Further Reading:--


Together with quantum computers,quantum mechanics and Quantum mathematics will change the future of Artificial Intelligence.

In current computation cost and limitation the super invention complex number usage is limited,many statistical problems and algorithms are in queue waiting to process and make it in production,Quantum computers are not able to solve it as the current computation error is high,Quantum mathematics won't die and special computation logic will come to tackle this ,More info available


Not the answer you're looking for? Browse other questions tagged or ask your own question.