Google is making its robots smarter by teaching them to think for themselves
Robots are getting smarter, which means they are better able to execute our commands. A number of different companies worldwide focus their attention on creating robots but one company in particular is really taking the lead on this lofty goal: Google.
A new language model
According to Fast Company, Google is achieving this feat through its latest milestone in robot software research - a new language model called PaLM-SayCan (Pathways Language Model). The software has been engineered in collaboration with Everyday Robots and it provides the company's robots with an understanding of more complicated human requests that require comprehending and executing spoken commands.
To accomplish these requests the robots must break them down into smaller steps or goals and use the corresponding skills they have for them. So far, Google says PaLM-SayCan makes robots 14 percent better at planning jobs and 13 percent better at completing them. The firm has also reported a 26 percent improvement in robots’ ability to plan tasks involving eight or more steps or goals.
However, Google Research robotics lead Vincent Vanhoucke told Fast Company that these advanced robot models are all about exploring what the company can do. “Google tries to be a company that focuses on providing access to information, helping people task in their daily lives,” he explained.
“You could imagine a ton of overlap between Google’s overarching mission and what we’re doing in terms of more concrete goals. I think we’re really at the level of providing capabilities, and trying to understand what capabilities we can provide. It’s still a quest of ‘what are the things that the robot can do? And can we broaden our imagination about what’s possible?'"
Robotics and ping pong
Vanhoucke also told TechCrunch that the current issues in robotics can best be exemplified by the use of robotics in ping pong.
“You may wonder why ping-pong. One of the big challenges in robotics today is this intersection of being fast, precise and adaptive. You can be fast and not adaptive at all; that’s not a problem. That’s fine in an industrial setting. But being fast and adaptive and precise is a really big challenge. Ping-pong is a really nice microcosm of the problem. It requires precision and speed. You can learn from people playing: it’s a skill that people develop by practicing,” Vanhoucke added.
“It’s not a skill where you can read the rules and become a champion overnight. You have to really practice it.”
Although Google's robot may not have mastered the sport, it does understand other languages and the team behind the machines says the language skills were free. They demonstrated this by asking the robot for a Coca Cola in French. What does this mean for the development of more advanced robotics?
For now, Google is focused on making commands more understandable and perhaps relatable to robots and that is an excellent skillset to develop in what may be our future assistants. The better they make them at understanding our commands and executing them, the more popular they are likely to be. Are you ready for the future of robotics?