Our researchers are leading the way in revolutionary advancements like using quantum-enhanced artificial intelligence (AI). This creative new method aims to address some of mankind’s greatest challenges, from climate change to disease. This creative strategy has the potential to spur solutions to all these problems years before they would otherwise be addressed. Now, scientists are using the power of quantum computing. Yet their ultimate goal is to prove out useful, meaningful chemistry simulations that classical computers will be unable to reach.
In order to realize such breakthroughs, we must create hundreds to thousands of qubits of excellent quality. Additionally, these qubits must have error rates on the order of 10^-15—one error allowed in a quadrillion operations. Practitioners are already working on ways to train AI on quantum-accurate data. This advance could enable faster predictions and assessments in virtually every scientific discipline.
The Power of Quantum Computing and AI
Quantum computing and AI are emerging as potentially transformative tools in developing new materials science and chemistry. With traditional computational approaches this often proves intractable, especially for simulating molecules that have more than a few hundred atoms. This constraint requires considerable computing power and time, often covering days or weeks for simulations. Quantum computing promises the opportunity to deliver a much more precise simulation of such complex systems.
Getting to even the level of reliability that’s needed for quantum computing is a tall order. Researchers must embed fault tolerance by redundantly encoding quantum information into logical qubits. Each of these logical qubits consist of hundreds of physical qubits. To achieve the error rates we want, we might need upwards of a million physical qubits altogether.
The good news is that as scientists work to improve these technologies, they’re beginning to pinpoint the most promising paths for applying AI. One example of this is a way in which these models are being used — development of new battery materials. Our advanced AI models are working around the clock to rapidly screen more than 32 million candidates for new battery materials. This groundbreaking technology allows researchers to identify safer, cheaper, more effective and sustainable alternatives.
Speeding Up Material Discovery
The U.S. Pacific Northwest National Laboratory (PNNL) partners with several academic institutes. Together, their amazing work highlights the transformative potential of quantum-enhanced AI. In under a week, researchers narrowed that original 32 million potential battery materials down to 500,000 stable candidates. They ended up pinning down 800 that had amazing potential. This breakthrough is a remarkable demonstration of the efficiency gains AI can make. Using traditional methods it would have taken nearly two decades to interview this large a group of potential candidates.
Using AI specifically trained on quantum-accurate data, researchers can direct their finite resources toward the most promising candidates with exponentially greater efficiency. Quantum-enhanced AI will soon help reduce the many costs of laboratory synthesis and testing by orders of magnitude. It helps researchers focus their efforts on “first-time right” candidates—the ones predicted to work in the first try, according to more sophisticated models.
“This collaboration demonstrates how we can leverage AI and quantum computing to expedite the discovery process for new materials,” said a PNNL researcher involved in the project.
It is critically important to be able to accurately simulate electron correlation in systems where strong electron interactions are at play. This understanding allows us to create materials with exotic electronic behaviors, such as high-temperature superconductors. Quantum computing will allow us to scale this kind of detail, enlightening our understanding and potentially creating breakthroughs into everything from new energy sources to biomedical advancements.
Future Implications for Scientific Research
As we move forward with quantum-enhanced AI, the impact it will have on scientific research in nearly every field is apparent. Traditional simulations have held science back. Now, as with much of the research landscape, AI is revolutionizing the field, accelerating research drastically and opening up new possibilities while replacing these labor-intensive strategies.
Moving from strictly classical computation to leveraging quantum-enhanced techniques is an exciting change in the way researchers will be able to tackle their most complex problems. Scientists are learning to exploit the potential of quantum computing to vastly increase their capacity to run simulations. Along the way, they’re improving their core knowledge of complex systems.