Artificial intelligence comes to quantum computers

 In Computing, Cutting edge technology

Quantum artificial intelligence

Spanish researchers have developed artificial intelligence algorithms to be run on quantum computers.

In a work that combines physics, biology, and quantum computing, the team used simulators – since powerful enough quantum computers are not yet available – to create evolutionary algorithms that mimic life, natural selection, learning, and memory.

In fact, these algorithms are essential for the very development of quantum computers, which need reliable programs so that their functioning can be attested – whoever accompanies the development of that field is certainly reminiscent of Shor’s algorithm, a quantum factorization algorithm that has become A key piece for the creation of quantum processors themselves.

As the new algorithms reproduce in certain quantum systems exclusive properties of living entities, Unai Alvarez Rodriguez and his colleagues at the University of the Basque Country coined a new term to describe his work: quantum biomimetics.


Quantum Biomimetics

The first algorithm recreates a natural selection environment in which qubits function as individuals that replicate, mutate, interact with others and the environment, and even reach a state equivalent to death.

“We developed this final mechanism for individuals to have a finite lifespan,” explained the researcher. Thus, in combining all elements, the system does not have a single clear solution: “We approach the natural selection process as a dispute between different strategies in which each one individual would be a strategy to solve the problem. And the solution would be the strategy capable of Master the available space,” Rodriguez said.

A quantum simulator essentially allows you to “pilot” atoms to see how quantum particles behave. [Image: Britton / NIST]

The algorithm to simulate memory, on the other hand, consists of a system governed by equations. These equations have a dependence on their previous and future states, so the way the system changes “does not just depend on how it’s at the moment, but where it was 5 minutes ago and where it’s going to be in 5 minutes,” Rodriguez explained.

Finally, in the quantum algorithm related to the machine learning process, mechanisms were developed to optimize well-defined tasks, to improve the classic algorithms and to improve the margins of error and the reliability of the operations.

An unexpected result was that “we were able to encode a function in a quantum system, but not to write it directly, the system did it autonomously, we could say that it ‘learned’ through the mechanism designed to make it happen. New advances in this research, “said the researcher.

From computational models to the real world

“Although they have been done theoretically, we propose simulations to work in experiments on different types of quantum platforms, such as ion traps, superconducting circuits, and photonic waveguides, among others. , “Finished Rodriguez.

If quantum computers are expected to show gains in speed, the possibility of directly simulating biological processes in artificial intelligence and machine learning systems poses new computing possibilities whose scope is difficult to predict.

Recommended Posts

Leave a Comment


Start typing and press Enter to search