The quantum computer is already obsolete in the face of this disruptive technology which is overtaking it in the scientific field

Could Artificial Intelligence outperform quantum computing?

As tech companies invest of the billion in quantum computershoping to revolutionize fields as varied as finance, drug discovery and logistics, rapid advances in artificial intelligence (AI) in physics and chemistry simulations are leading some to wonder if we will even need quantum computers.

Read also:

The competitive rise of AI

As the field of quantum computing grapples with the realities of fickle quantum hardware, AI is making significant advances, challenging the supremacy of quantum computers, especially in the areas of fundamental physics, chemistry and materials sciences. The capacity and complexity of quantum systems that AI can simulate is rapidly advancing, which could shift the supposed playing field for quantum computers.

This Pacific island succeeds where the United States and Europe failed with this extraordinary rocket plane: carrying out 2 flights with the same aircraft in 8 hours

Power and limitations of quantum computers

Quantum computers promise to perform certain calculations much faster than classical computers, a promise that would, however, require much larger quantum processors than those available today. Quantum algorithms, such as Shor's algorithm, could solve problems exponentially faster than classical algorithms, but only if the problems involved allow full exploitation of quantum effects.

The barriers of quantum computing

A recent study co-authored by Matthias Troyer, head of quantum computing at Microsoft, found that the theoretical advantages of quantum computers diminish when taking into account the relative slowness of quantum hardware compared to modern computer chips. Additionally, the difficulty of integrating large amounts of classical data into a quantum computer represents a major obstacle.

Benefits of AI in simulating weakly correlated systems

AI has proven that it can simulate weakly correlated quantum systems with high efficiency using classical tools like density functional theory (DFT). These systems are simpler to model because interactions between particles are minimal, making quantum computers less necessary for these types of problems.

Contributions of AI to chemistry and materials sciences

Recent advances in using AI to generate data on chemicals, biomolecules and materials, which is then used to train neural networks, are revolutionizing the ability to predict the properties of chemical structures. These AI models, which learn patterns in the data, are much less expensive to run than conventional DFT calculations.

Perspectives futures

Future research could enable AI to simulate even the largest weakly correlated systems, making quantum computers potentially obsolete for these applications before they even become widely available. At the same time, highly correlated systems, which remain a challenge for DFT, are also beginning to be accessible thanks to advances in AI, notably using neural networks to model the excited states of quantum systems.

The United Kingdom is preparing for the supreme sacrilege: disarming its only two aircraft carriers, which would leave France master of the European seas

This article explores how recent advances in artificial intelligence could threaten to make quantum computing obsolete in many fields, particularly those where quantum effects are dominant. While quantum computers still struggle with significant technical limitations, AI is rapidly advancing in the simulation of complex systems, raising the question of whether quantum computers are truly necessary in the future of research and industry.

Source: MIT

Post Views: 18