VW To Use Google Quantum Computer For Battery Research, AI
Volkswagen announced that it will use Google’s universal quantum computer to research new battery technologies for its upcoming electric vehicles, make its self-driving systems smarter, and improve traffic flow in cities.
Volkswagen: An Early Adopter Of Quantum Computing
Volkswagen was the first carmaker to adopt quantum computers. Earlier this year, it announced its first successful project to significantly improve travel times for 10,000 taxis in Beijing using a D-Wave quantum annealing computer.
However, the company now seems set to try Google’s potentially more advanced universal quantum computer. D-Wave is a more specialized type of quantum computer, so Volkswagen should be able to develop more types of projects that can take advantage of quantum effects with Google’s more “general purpose” quantum computer.
“Volkswagen’s collaboration with Google marks the beginning of quantum computing in the automotive industry, and is a paramount step to addressing modern mobility challenges unlikely to be solved with binary digital electronic computers,” said Abdallah Shanti, Executive Vice President and Group Chief Information & Digital Officer for Region Americas, Volkswagen of America, Inc. “Through this partnership, Volkswagen intends to unlock the potential of this technology, and share our learnings to motivate the development of quantum computers and algorithms,” he added.
Volkswagen’s Collaboration With Google
Volkswagen wants to focus on three main areas of research when using Google’s quantum computer: high-performance batteries for electric cars, artificial intelligence, and traffic optimization.
Battery Research
We know quantum computers are supposed to be good at helping researchers find new materials, so using a quantum computer to find or develop a better battery technology makes sense. Electric vehicles may be on their way to revolutionize the car industry, but for now the high prices for their batteries contribute to slow electric vehicle adoption. The faster batteries drop in price, the faster electric vehicles will be adopted.
Stay On the Cutting Edge: Get the Tom's Hardware Newsletter
Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.
Machine Learning
It’s no surprise that Volkswagen would want to research advanced AI, as that is also going to be necessary for building safe autonomous cars. Quantum computers should be able to make machine learning much more efficient.
Right now, if you want to train an AI agent that has a high accuracy in identifying street objects, for instance, you will need to throw millions and millions of images of such objects at it. However, with quantum computers, you should be able to use much smaller data sets to reach the same level of accuracy.
Optimized Traffic Flow
Volkswagen has already used D-Wave’s quantum annealing computer to develop advanced traffic optimization algorithms, but it seems the company believes Google’s universal quantum computer could take that project further. Volkswagen plans to develop algorithms for urban traffic guidance systems, as well as identifying available electric charging stations or vacant parking spaces.
Even though we haven’t reached “quantum supremacy” yet -- the point in time when quantum computers are faster than the fastest classical supercomputer -- it looks like the quantum computing revolution is already sneaking up on us. As more big companies start to either develop quantum computers or use them to improve their own technologies and businesses, the quantum computing field should start maturing more rapidly. This should result in more powerful and more useful quantum computer that can be used by even more companies and academics to increase the rate of research.
-
bit_user
Is this accurate? I thought the benefit of quantum computer for deep learning was to arrive at an optimal solution in a single run, rather than taking many iterations to converge on a sub-optimal "good" solution.20355909 said:Right now, if you want to train an AI agent that has a high accuracy in identifying street objects, for instance, you will need to throw millions and millions of images of such objects at it. However, with quantum computers, you should be able to use much smaller data sets to reach the same level of accuracy.
Or, perhaps you're thinking that the optimal solution will require fewer nodes for a given accuracy requirement, meaning less training data will be required to mitigate against over-fitting?
-
milosz VW should use powerful AI to learn how to build reliable cars and provide quality service.Reply