Two of the hottest areas of technology today are artificial intelligence and quantum computing.
Not only are these technologies generating excitement in the scientific community, but both have captured the imagination of the public. Indeed, hardly a day passes when the media does not feature a prominent story about AI – sometimes about how wonderful its potential will benefit everyone’s lives, and sometimes about how it is on the verge of destroying any prospect of future employment and of taking over everyone’s lives in a “Big Brother” scenario. At the same time, although what quantum computing is, and how quantum computing works, are not understood by most – even technical people – quantum computing is generally viewed by the public as a fantasy of the future. To be sure, AI and quantum computing are fundamentally different technologies. They are, however, inextricably intertwined.
Although “artificial intelligence” may appear to be fanciful, it really is nothing more than an ordinary solution to a typical engineering problem. It starts with modeling a problem and adds an algorithm for solving that problem. The problem – and the solution – may have increasingly complicated aspects. Take, for example, machine vision, which is basically an algorithm to enable a computer to have the artificial intelligence to “see” something. An example of a simple problem is enabling the computer to locate the position of an object by detecting edges or corners of that object. From there, machine vision can become substantially more complicated in application. It can look at an object that has been manufactured and compare that object to the desired template shape to see if the manufactured object is defective; it can detect an object and compare it to shapes in memory to determine what that object is (e.g., is it a cat or a car?); it can make an image of a person’s face and identify that person from face data stored in memory; it can make an image of an older person’s face and attempt to identify if that person is the same person whose younger face data is stored in memory.
As artificial intelligence applications become increasingly sophisticated, the required algorithms create increasing computational demand.
Larger data sets must be analyzed with increasingly complicated calculations and feedback. Artificial intelligence – like human intelligence – does not review information to arrive at the correct answer to a question. Rather, by evaluating current information and comparing that information to their past experience, people reach a reasoned judgment about what may be the best answer to a question. If that answer turns out to be inadequate, they repeat the process to find a better solution. Computers do the same thing. A computer reviews data and engages in “machine learning” from past information to arrive at an optimized solution. That process can place a heavy demand on computational capacity.
That is where quantum computing comes in. Quantum computers have a substantially greater computational capacity than conventional computers. It has been said that a quantum computer may be able to perform more calculations at the same time as there are atoms in the universe. Thus, quantum computers can execute billions of calculations simultaneously. As a result, the computationally intensive requirements of sophisticated artificial intelligence applications, which are beyond the capability of conventional computers, may be met by quantum computers.
In this regard, quantum computing may follow a similar path to the path digital recording has followed over the last 45 years. At the start of that time, sound recordings were made by recording analog signals and images were made using chemical photography. The engineering of how to use digital means to record such information was well understood then, and viewed with excitement. However, the processing capacity available at the time could not generate clear sounds and images. As computational capacity expanded, digital recording and imaging have become nearly universal today. Indeed, they are generally superior in quality to the quality of the old analog and chemical methods. One can likewise expect quantum computing to be used increasingly over time for artificial intelligence applications that cannot be performed – or cannot be performed well – today. There are numerous applications that stand to benefit from the combination of artificial intelligence and quantum computing, such as development of new pharmaceutical products, cryptography, detection of fraud, analysis of financial opportunities for investment, optimizing manufacturing systems and translation between languages.
The synergistic relationship between artificial intelligence and quantum computers is not a one-way street.
Plainly, the increased capacity of quantum computing will support new, improved and increasingly complex artificial intelligence applications. But it is also true that, in return, artificial intelligence is available to assist in the development of quantum technology, which relies on understanding and modeling the complex interactions that occur between quantum particles and waves.
Perhaps it would be better said that the relationship between artificial intelligence and quantum computing is a “match made in heaven.”