The California Institute of Technology today announced that Amazon Web Services Inc. is opening a new research hub on the university’s campus to develop cutting-edge quantum computing technology. 

The research hub, called the AWS Center for Quantum Computing, is reportedly opening this week. It’s located inside a two-story facility in the northeast corner of the Caltech campus that is described as the university’s first “corporate-partnership building.”

The Washington Post reported that research activities at the AWS Center for Quantum Computing will be led by Caltech professors Oskar Painter and Fernando Brandão. The two professors also hold senior roles at AWS. Painter is the cloud giant’s head of quantum hardware, while Brandão is the head of quantum algorithms.

The cloud giant has reportedly recruited scientists from multiple universities to support the center’s work. The intellectual property that the research hub produces will be owned by Amazon, the Post reported. Separately, the company is sponsoring a number of research projects at Caltech, and the rights to the innovations produced through those initiatives belong to the university. There are some cases where Amazon and Caltech will share intellectual property.

The reason Amazon and other tech giants are investing in quantum computing is that the technology holds the potential to solve calculations too complex even for the most powerful conventional supercomputers. It’s believed that tomorrow’s quantum machines could help advance research efforts in a variety of scientific fields. The technology may also lends itself to accelerating machine learning algorithms.

For Amazon, quantum computers’ potential use as a supply chain optimization tool could prove particularly valuable. Finding the most efficient way to transport packages to customers or between warehouses requires evaluating a massive number of potential routes, as well as numerous other factors such as weather conditions.

Tomorrow’s quantum computers could excel at this task. That’s because qubits, the digital equivalent of bits, can perform calculations simultaneously instead of one after another like transistors, which speeds up processing. 

But before quantum computers can be applied to tasks such as discovering optimal package delivery routes, they have to be scaled up. “We can do small problems now with quantum computers but we need to scale up the technology by many orders of magnitude before we can truly tackle problems with a big impact,” explained Painter, the Caltech professor and AWS quantum hardware head.

Currently, one of the main obstacles to building large-scale quantum computers is error correction. 

To run complex algorithms, a computer first and foremost requires the ability to carry out calculations accurately. Today’s quantum computers have only limited accuracy because quantum circuits are highly prone to errors. Researchers are developing specialized  error correction techniques to increase the reliability with which calculations can be performed and thereby facilitate the development of more powerful quantum machines. 

“Transistors in our modern computers experience extremely low error rates at the level of one error per billion billion operations, enabling complex computations to be performed,” said Painter. “Quantum computers as of now are limited by error rates at the level of approximately one error in every thousand operations.”

In traditional computers, the usual approach to filtering data errors is to create multiple copies of information. If an issue emerges in one of the copies, the others can be used instead. But that technique is not available to quantum computers, which can’t create multiple copies of the information they process because of a quantum mechanical phenomenon known as the no-cloning theorem. Scientists must find other, creative ways of detecting and fixing data errors.

The most common quantum error correction techniques involve creating groups of qubits that each function as a single “logical qubit.” The qubits in a given group work together to identify and fix potential mistakes that could affect the reliability of calculations. But though the technology has become more sophisticated in recent years, it’s still far from perfect. 

Today, carrying out quantum error correction is so computationally intensive that the process often takes up a sizable percentage of quantum computers’ processing power. Moreover, the technology can only resolve certain types of calculation mistakes and, in the process, often introduces new errors into the mix.

“Going forward, we want to scale up the number of logical qubits to hundreds or thousands, while also driving the logical qubit error rate down by multiple orders of magnitude, so that we can perform quantum computations of sufficient complexity to tackle high-value problems,” said Painter. “In order to do so, we need to further develop both the physical hardware and the logical qubit architecture.”

AWS already provides quantum computing infrastructure through its public cloud. The Amazon unit offers a service called Amazon Braket that enables organizations to access quantum computers from Rigetti Computing Inc., IonQ Inc. and D-Wave Systems Inc., three startups that are each using different qubit technologies. 

Because AWS reportedly owns the rights to the innovations produced by the new AWS Center for Quantum Computing, the research being carried out at the hub could in the long term help the cloud giant enhance its Braket service. That’s particularly significant given how AWS’ rivals in the public cloud are also developing their own quantum computing technology. Microsoft offers access to quantum computing hardware and software tools through its Azure cloud  platform, while Google LLC earlier this year set a goal of developing an error-free quantum computer by 2029. 

Photo: Caltech