IonQ and Hyundai Team Up for Automotive Object Recognition
Signifying a step toward practical application of quantum computing technology, IonQ and Hyundai have partnered to apply quantum machine learning (ML) for object recognition of common roadway objects.
Quantum computing technology has quickly moved from theoretical to practical in recent years. It’s been used for a variety of compute-intensive tasks such as electric vehicle (EV) battery chemistry analysis and large-scale financial simulations.
Today quantum processing startup IonQ and Hyundai Motor Company announced a joint effort that centers on an application of quantum computing that will impact our day-to-day lives. The effort takes the form of a new project to leverage quantum ML to do image classification and 3D object detection for automotive-based platforms, especially for systems where such technology is vital: autonomous vehicles (AVs).
An example of object recognition of road signs. Image used courtesy of MDPI and Villalón-Sepúlveda et al
In this article, we’ll cover what the partnership involves, what kind of quantum processing is used, and what’s behind the new benchmark metrics for IonQ’s Aria quantum computer.
Quantum Computing for Heavy Data Loads
For this project, IonQ says that it will collaborate with Hyundai to use IonQ quantum computers to perform ML tasks more efficiently. Quantum computers process huge amounts of data and do it faster and more accurately than traditional computing methods, says the company. The company unveiled its IonQ QPU architecture last summer. It revolves around an approach it calls "Trapped-ion" quantum computing that uses ionized ytterbium atoms as qubits.
IonQ’s EGT chip utilizes “trapped-ion” technology. Image used courtesy of IonQ
IonQ brings to the effort work it’s already done encoding images into quantum states. Using that technology, the company says it can use its quantum processors to classify 43 types of road signs. In the next phase of the project, the two companies plan to apply IonQ’s ML data to Hyundai’s test environment.
While using ML to detect road signs is fundamentally a 2D process, the more complex and broader project is detecting 3D objects in an automotive context. From there, they plan to simulate various real-world scenarios. With that in mind, the two companies plan to leverage their existing work on road sign recognition and apply it to “3D” objects like cyclists and pedestrians.
InoQ Aria Quantum Computer Goes to Work
To do this quantum computing work, it’s expected that the companies will run object recognition tasks on the IonQ Aria quantum computer. There is little in the way of publicly available details about the IonQ Aria—no product page with detailed specs, for example. That said, the company has made claims that it is the industry's “most powerful quantum computer.” This claim is backed up by algorithmic qubits benchmark data that the company presented at last month’s Quantum Information Processing (QIP) conference.
At that event, IonQ introduced its new algorithmic qubits (#AQ) benchmark, a single-number benchmark for measuring quantum computers' performance. The #AQ metric was derived from algorithm benchmark protocols that were the result of an independent study done by the Quantum Economic Development Consortium (QED-C).
This shows the IonQ Aria quantum computer's small form factor “deck of cards” trap and vacuum chamber package. Image used courtesy of IonQ
The IonQ Aria has an #AQ of 20. The company says that means it can execute quantum circuits containing more than 550 gates, in contrast to superconductor-based systems that can only execute circuits comprised of dozens of gates. Which, they believe, gives the IonQ a leg up for practical uses of quantum algorithms.
Another Use Case for Quantum Computing
This announcement today isn’t the first collaboration between IonQ and Hyundai. In January, the firms partnered to marry IonQ quantum computers and Hyundai’s electric vehicle (EV) battery data to improve efficiency and lower the costs of EV batteries. Other collaborations announced previously by IonQ include a deal with GE Research to analyze risk mitigation and with Goldman Sachs involving simulation of financial functions such as pricing options.
Meanwhile, companies like Sandbox AQ see a future for quantum processing in critical compute-intensive applications like cryptography and cybersecurity. But this new deal between IonQ and Hyundai represents a use case for quantum processing that could become an enabler of systems involving things we see—or driverless cars see—right in front of our eyes.