Panasonic is joining with a Tokyo-based venture to help its driver-assistance computers learn on the job, part of efforts in Japan to join the “deep learning” vogue that has swept Silicon Valley.
People familiar with the deal said Preferred Networks would provide software to beef up Panasonic’s offerings in advanced driver assistance systems. The Osaka-based electronics giant hopes such systems and other auto-safety businesses will bring in some $3.8 billion a year in revenue by the fiscal year starting in 2018.
“Deep learning” is a kind of self-study program for machines. Instead of being programmed solely with fixed rules, computers would learn by themselves how to achieve a task—such as avoiding obstacles, in the case of cars. Eventually, all car computers could be connected to a single “brain.”
Panasonic hopes to leverage the venture’s deep-learning expertise in other products and services, including consumer electronics and surveillance cameras, the people familiar with the deal said.
Source: Wall Street Journal