Created AI, able to accelerate the development of new batteries by 98%

The development of high-density batteries is a difficult task for science – it takes too much time to evaluate the effectiveness of new chemical compounds. Although developers already have several promising projects, years or even decades can pass before their implementation.

Soon the situation may change – a team of scientists has developed a methodology that can reduce the time required for testing battery technology by 98%.

The new technique is based on the use of AI and machine learning. A team of researchers at Stanford University, led by professors Stefano Hermon and William Chue, has developed an algorithm that reduces testing time for new batteries by 98 percent. The effectiveness of the methodology was successfully tested in the search for the best way to charge an electric car in 10 minutes, which will extend the battery life. Thanks to the use of the new algorithm, the experiment time was reduced from 2 years to 16 days.

“When testing the battery, you need to try a lot of things, because the performance that you get in different conditions will be very different. With the help of AI, we can quickly determine the most promising approaches and eliminate many unnecessary experiments, ”the researchers note.

Scientists said that artificial intelligence can accelerate almost every stage of battery development: from analyzing the chemical composition to determining its size and shape, as well as finding more advanced systems for production and storage. It is predicted that through the use of machine learning, the development of new technologies for the production of batteries can accelerate dozens of times.