It was about four years ago, in the back of an Uber driving him back from a conference, that the idea of using quantum computing to design OLED displays for smartphones and TVs started germinating in the mind of Michael Helander, the CEO and co-founder of materials design company OTI Lumionics.
Helander was sharing his ride with a particle physicist who doubled as a VC, and who was then an early investor in leading quantum computing company D-Wave. As you do in such circumstances, the pair were discussing quantum computing solutions capable of simulating the properties of atoms coming together to form molecules and solids – and what that might mean for Helander’s field of expertise, computational chemistry.
“That conversation got me asking myself: is this even feasible?” Helander tells ZDNet. Now a few years later, it would seem so. OTI has successfully developed a new electrode material that is ready for mass production and started shipping worldwide at the end of 2020. The material will be used to manufacture first-of-their-kind transparent OLED displays.
Most OLED displays require several layers made up of different materials to function, including a cathode, through which electrical current flows in. Because standard cathodes are not transparent, front-facing cameras and sensors for technologies like facial recognition have to sit on top of the display, which is why most smartphones still come with a punch-hole at the top.
SEE: Tableau business analytics platform: A cheat sheet (free PDF download) (TechRepublic)
For our devices’ bulky cutouts to disappear, cameras would have to be integrated under the display – meaning that the display needs to be transparent. OTI’s team replaced standard cathodes with a new material patterned with small holes that act as microscopic transparent windows, effectively letting light go through the display.
With front-facing cameras and 3D facial recognition sensors moved under the display, not only can the screen be larger and smoother, but transparent displays also come with higher brightness and longer battery life. Helander hopes this will bring about new designs for phones, and also laptops, tablets and foldable devices, as well as AR and VR hardware.
“OLED displays are a massive and growing market,” says Helander. “There is a lot of excitement about the technology expanding into laptops and monitors. We see it as an opportunity to innovate when it comes to the convergence of display and sensors.”
Behind OTI’s innovative product is a so-called “materials discovery platform” – and powering that platform, equally as innovative techniques. “At OTI Lumionics we are developing advanced materials – by design – using quantum simulations, machine learning and real-world testing in pilot production,” proudly states the company’s pitch.
There is a good reason that Helander’s interest in quantum was piqued four years ago: the technology, although still in its infancy, is expected to break new ground in the field of molecular simulation. For the CEO of a company that describes itself as a designer of advanced materials for the electronics sector, that is enough to justify digging deeper.
From early on, Helander’s strategy has consisted of using a computer-based approach to electronic material design. As a small company, OTI was never equipped with armies of chemists ready to test and trial thousands of different molecular designs in the lab until a winning combination was found. “The way we develop materials has been heavily based on the use of computational techniques in chemical and material design,” explains Helander.
“But it turns out that even state-of-the-art classical computational chemistry, for a lot of these difficult problems, is inadequate,” he continues. “Either they can’t reach a high enough level of accuracy, or, if the theory is accurate enough, it becomes an intractable problem that requires a supercomputer to solve.”
Quantum computing, and its ability to leverage the odd behavior of qubits to solve many calculations at once, seemed at first glance an ideal match. Qubits could be used to predict how the complex alignment of many different compounds could result in particular properties for a given electronic material, as well as how this material would interact with other molecules in a device – and they could, in principle, do this faster and more accurately than any existing classical methods.
Around the same time, long-established quantum champion IBM published the results of an experiment showing that simple molecules like hydrogen could be simulated by a universal gate-based quantum system. The stars were aligned; the odds were in favor of quantum-based molecular simulation; and OTI’s chemists started getting excited about the implications for computational chemistry.
They quickly found themselves facing a limiting factor. With less than a hundred qubits currently sitting in most quantum computers, there wasn’t much that could actually be done. “To solve an industrial-sized problem, you need more qubits than will be scientifically feasible in the next ten to 20 years,” says Helander. “But as a small company, we don’t have the resources to invest in a long R&D program of that kind.”
Like any CEO, Helander’s interest lies in short-to-near-term business value; and so, he decided to tackle the problem with an entirely new perspective. If the number of qubits available couldn’t match the size of the problem, then the problem had to be re-made to match the number of qubits at hand.
“That’s actually a gap in theory,” says Helander. “So I started with a group of theoreticians. I told them to forget everything they knew about computational chemistry, and imagine a new set of computational chemistry representation to map to a qubit space. What would that look like?”
There is a long-standing problem in the quantum space, argues Helander: instead of developing brand-new programs that are tailored for quantum hardware, scientists apply classical models to qubits. As it turns out, however, the way problems are represented in the classical world doesn’t always sit well with small-scale, hardware-constrained quantum computers.
Take the unitary coupled cluster – that is, chemists’ jargon to describe the technique used to represent chemical systems. According to Helander, that particular classical representation is highly inefficient when mapped onto a quantum computer, and requires large numbers of qubits and gate operations. Instead, OTI’s researchers developed a brand-new “qubit coupled cluster method,” adapted specially for quantum systems.
“In order to see value with limited hardware, you have to develop native code and write low-level stuff,” says Helander. “We developed that first native representation of the problem we wanted to solve, for quantum computers.”
Theory was promptly built into software and, equipped with a bunch of new quantum-ready algorithms, OTI’s team tested the technology in cloud-based quantum computers. The researchers, however, couldn’t let go of an ongoing feeling of frustration – at the nevertheless limited hardware, at the lack of error correction, at the stubborn levels of noise, and often at all three at the same time.
This is when Helander started looking closer at quantum-inspired techniques, a branch of the field that looks at ways to apply quantum-optimized algorithms to classical hardware. With a new set of custom-built, highly efficient quantum algorithms, wondered the CEO, why not try and run the software on regular CPUs and GPUs?
A partnership with Microsoft soon followed, and OTI’s team started using the Redmond giant’s Azure Quantum platform, which is designed to run quantum-inspired algorithms on classical Azure hardware. In principle, by using sophisticated optimization techniques, Azure Quantum enables users to reap the rewards of quantum computing approaches while using classical devices.
Last year, in a blog post, Microsoft announced that the project was showing signs of success: OTI had effectively demonstrated meaningful results on commercially relevant sized problems. Specifically, the company had completed the simulation of a green light-emitting OLED material known as Alq3 – a problem that would have required 42 error-corrected qubits on gate-based quantum hardware.
For Helander, the experiment showed the promise of much nearer-term value to be drawn from quantum-inspired algorithms, and their potential to start drawing benefits from quantum computers without needing to use them directly.
That is not to say that OTI has ruled out using pure quantum hardware. Quite the opposite: the company is working with D-Wave, which provides a cloud-based quantum annealer that is much easier to control than the gate-based quantum computers operated by companies like IBM or Rigetti. This means that D-Wave can offer a technology that is already several thousands of qubits-strong, and that can reach the industrial relevance that Helander and his team are looking for, without error.
Helander and his team, therefore, share their time between classical techniques, quantum-inspired approaches and purely quantum-based experiments.
“At the moment, our quantum techniques focus a lot on theory development and optimization,” says Helander. “For our current product, for example, we applied a combination of all the different tools that we had – classical simulations, quantum systems and quantum-inspired algorithms.”
“We still heavily combine our quantum methods with classical techniques,” he continues. “Even though the amount of value we are driving is only a small subset of our everyday work, from this point forward we’re looking at increasing that over time until more of our workflow is adopting quantum and quantum-inspired methods.”
While the company, for now, is focusing on high-value OLED displays, Helander is positive that the discoveries led by OTI’s research team will generate an avalanche of innovations in many other fields such as battery design and drug development. The technology could effectively replace processes that were until now based on trial-and-error, with highly sophisticated computer models that would rapidly build designs for new molecules from the ground up.
The potential of quantum computing to phenomenally disrupt industries that are hunting for new and improved materials is well-known, but it will be at least a decade before quantum’s value translates into real-world results. For those too impatient to wait, however, quantum-inspired methods might provide an early sneak peak of better things to come.
This is a syndicated post. Read the original post at Source link .