1
$\begingroup$

This question already has an answer here:

According to the press information from Dwave, quantum annealing is up and running and offers a lot of opportunities for society. “Basically, a quantum computer is a non-deterministic turing machine which is exciting and cool” – this is at least the statement of AI influencer Siraj Raval who has researched the topic in detail. He has made a video about the subject and is trying to communicte the subject to larger audience. Dwave itself is also interested in promoting their product. They have a lot of information out there and some companies are using these machines, for example Google.

But what i didn't understand is, why exactly do society needs a quantum computer? I mean, image recognition and robotics control can be done with classical deeplearning on mainstream hardware, why do we need superposition and qubits. Or asking more directly, which challenges are out there, that only quantum computers are able to master it?

$\endgroup$

marked as duplicate by nbro, DukeZhou Jan 15 at 20:11

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • $\begingroup$ How is this question related to AI? Maybe you should have asked it here: quantumcomputing.stackexchange.com. If you're asking how can QC be used to develop AI, then there's already another question: ai.stackexchange.com/q/36/2444. $\endgroup$ – nbro Jan 15 at 17:09
  • $\begingroup$ Provisionally closing because this question is more concisely stated as "What types of problems are quantum computers more suited for than classical computers?", which does seem to have an answer here: ai.stackexchange.com/questions/36/… $\endgroup$ – DukeZhou Jan 15 at 20:11
1
$\begingroup$

During the first decade of this century (and before that) buying a new computer after 2-3 years used to come with much better features and performance, in terms of processor, than it is now. We all know i7 and i5 are not that much different when it comes to performance for a standard home desktop.
As we all know that more the most of the modern technological era of human history the computational power and speed has been doubling for almost every two years (18 months). This observation has a name, known as Moore's Law. You can go over this Wikipedia link of Intel's Processors comparison. But now we are on the edge of compacting more and more transistors per unit area. The limitation is due the several reasons, like heating effect caused (leading to more power loss), quantum tunnelling effect.
Now on to Quantum Computers
We can't really compare quantum computers with the classical ones like we compare two cars of different models. Comparing both of them is more like comparing a bicycle with Boeing. QCs belong to the different generation of computers.
So, lets talk about the limitations of classical computers specially in terms of AI and ML:
We know that most of ML and a great part of AI is about mathematical computations. Classical computers are not good in solving problems when it involves solving factorial and/or exponential problems. These kind of problems mostly falls in the category of NP Complexity. To understand more about complexity you can go over this awesome thread in CS StackExchange. Another example will be of Curse of dimensionality. This problem is caused while analyzing data in hyper-dimensions.
Now onto QCs
Quantum Computers are fundamentally different. Instead of using bits (used by classical computers to store either 1 or 0) to store data they use quits that can be in the superposition of any state between 0 and 1. To understand more about this you can go over my Quora answer where I explained workings of QCs corresponding to the Shor Algorithm. Not only that but the gates that are used in QCs are different as they are reversible in nature. By reversible I mean knowing the output state you can determine the input values for the gates. Except NOT gate other gates are not reversible in classical computers.
Apart from all this there is no way we can use classical computers for molecular modelling, simulating bio-chemicals, DNA, etc. QCs definitely add efficiency and speed to the things done in terms of mathematical computation. Apart from just providing efficiency and speed, we need QCs just to get many things done, which are not possible with classical computers.

$\endgroup$
1
$\begingroup$

Short version:

Most likely we need them because as our need for faster calculations in less time and higher security rises with the increasing usage of internet, technology and A.I., we need faster and more powerful ways of computing. Quantum computers are in theory much faster and more powerful then conventional computers and also almost "impossible" to hack. Because of the resources and security Quantum computers provide spending resources on improving them could be a logical choice for us considering what we know today regarding our rising need for more computing resources.

$\endgroup$
0
$\begingroup$

One killer application is simulating quantum dynamical systems, which we need to do if we want to improve our (horrifically expensive and failure prone) pharmaceutical development pipeline, and there are applications in improving industrial chemistry (important processes for things like making nitrogen fertilizers are horrendously energy intensive) and in material science (think carbon-nanotube-based materials that make steel look soft as copper).

Yet to be fully worked out, there may be some huge payoffs for machine learning, which can be framed as finding the energy minima of a system, something that quantum dynamical systems can do with beyond-classical efficiency.

$\endgroup$

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