Will Quantum Computing Revolutionize AI?
AI and Quantum Computing therefore seem destined to coexist as distinct tools, often complementary, but not always compatible.
In recent years, quantum computers have raised hopes of profoundly transforming the fields of innovation and research. From cryptography to the challenges of superconductivity, not to mention drug design, these computers of tomorrow, powered by the quantum properties of matter, offer the possibility of breakthroughs. The reason? Their computing power is far greater than that of conventional computers.
Among the applications envisaged, AI appears to be a particularly promising field. But are these expectations realistic? Are we on the verge of an AI revolution thanks to Quantum Computing?
Large volumes of data
Since around 2018, scientists have been predicting that the incredible computing power of Quantum Computing could accelerate the development of AI, particularly in deep learning. This field of AI is driving new technologies, such as generative models for text (like ChatGPT), audio, and video.
What fuels this conviction? The idea is that quantum computers would enable AI to operate more efficiently by establishing connections between data at unprecedented speed. Several researchers have therefore set out to reconcile these two technologies. This is an exciting prospect, reinforced by announcements from companies promising even more powerful quantum computers in the years to come.
Unfortunately, even if quantum computers have great information-processing power, they struggle to handle the huge volumes of data required by AI. Indeed, to generate text, images, or other content, AI requires the exploitation of gigantic quantities of data.
While quantum computers have enormous computing power, they are still very slow when it comes to inputting and outputting this data. For example, a quantum computer of 2030 could have a read/write speed comparable to that of a classical computer of the late 1990s. What's more, these computers can only perform short calculations without failure: it would take at least another fifteen years for them to become fault-tolerant.
Given this, should we abandon the possibility of a relationship between Quantum Computing and AI?
Not necessarily, because AI could come to the rescue of quantum computers. Indeed, AIs that work through reinforcement learning, where an algorithm adjusts its actions according to its environment and experiences, can drive these machines of the future forward. These machines are still in the prototype stage, and their characteristics vary widely from one model to another. This is why reinforcement learning algorithms could adapt to the specific features of each one and help them to work better.
What's more, when coupled together, Quantum Computing and AI could deliver faster, more efficient calculations in fields such as molecular simulation. In this field, classical computer algorithms are limited because they consume a lot of energy and can't work with molecules that are too complicated. It is by replacing them that Quantum Computing could shine with more efficient results.
AI and Quantum Computing therefore seem destined to coexist as distinct tools, often complementary, but not always compatible. More and more research teams at the intersection of physics, AI, and quantum simulation are collaborating. This type of collaboration embodies today's scientific effervescence and testifies to the potential for shaping tomorrow's innovation.
Could the Virtuous Circle of "Bitcoin Treasury Company" That Michael J. Saylor Is Selling to Investors Turn Into a Destructive Vicious Circle?
Warning, do not read this article if you are a blind Michael J. Saylor fan.