DeepMind, an artificial intelligence company based in London that Google bought for £400 million ($486 million) in 2014, is deeply in debt and owes its parent company, Alphabet, upwards of £1.04 billion ($1.26 billion), according to a report by the Financial Times.
The company’s losses also grew from £302.2 million ($367 million) to £470.2 million ($571 million) over the course of 2017. Even though DeepMind’s revenues have increased almost double from 2017, to £102.8 million ($125 million), the company has been weighed down by staff costs.
DeepMind’s staff costs doubled to £398 million ($483 million) last year, and they also doubled the previous year as well. So far, DeepMind’s only customer has been Alphabet, which compensated the firm £54.4 million ($66 million) in 2017.
Google confirmed in November that it was going to take over DeepMind Health, which employs about 100 people, but that hasn’t yet been finalized. The money owed doesn’t mean that DeepMind isn’t actively working on new projects, however.
Recently, DeepMind teamed up with Waymo to better train self-driving software in a way that mimics Darwinian evolution. Both companies are owned by Alphabet.
The companies use a method called population-based training (PBT), which lessens false positives when the software performs actions like placing boxes over the moving objects it sees in its sensors. The new training method also uses 50 percent less time and resources than previous methods.
PBT aims to help systems learn more efficiently. The neural nets in the software try an action, and then measure it against a previous action to determine whether it was more “right” or “wrong,” the report noted.
In previous methods, Waymo would have a lot of neural nets working by themselves on the same task, and with a different degree of deviation, or learning rate, in the way they approached the task, whether it was for something like identifying foreign objects or stopping in a timely manner.
A lower learning rate means steady progress and a higher learning rate means more variety in the quality of the outcome. The comparative training takes a lot of work to get right, because engineers have to either use a “gut feeling” to determine what’s right or manually search results and get rid of the badly performing ones.
PBT essentially automates this process, killing off the bad training and replacing it with better approaches, similar to evolutionary theory.