Hyundai Motor Company and the American startup IonQ have announced a new project in the field of quantum computing. The partners decided to use quantum machine learning to classify images, detect and recognize three—dimensional objects – all of this will be useful for future unmanned cars. The advantages of quantum algorithms in neural networks are obvious — it reduces the cost of computing and increases the speed of work by hundreds or even thousands of times.
Hyundai is already using IonQ quantum computers to calculate the structure and energy of lithium oxide — this is the first joint project of the companies, announced in January this year. An American startup builds quantum computers on ion traps that hold isolated ytterbium ions in a magnetic field, and, for example, its IonQ Aria on 20 algorithmic qubits is the most powerful in the industry.
Aria allows you to perform a reference quantum circuit with more than 400 valves and provides high accuracy of calculations. Using promising methods of encoding images into quantum states, IonQ has taught the neural network to classify 43 types of road signs, and now this data will be uploaded to the Hyundai test environment to simulate various driving scenarios in the real world.
In the future, there will be more sophisticated quantum algorithms capable of detecting and recognizing pedestrians and cyclists. All this will undoubtedly be useful to unmanned cars. However, quantum machine learning technology has other applications. For example, a quantum voice assistant will be ready to answer a question even before he hears the end of the phrase — and this will make him a truly full-fledged interlocutor.