Everyday Robot that learns from the world around it
Alphabet’s X research and development division revealed that they are working on the Everyday Robot Project. Basically, the goal of the company is to build a “general-purpose learning” robot. The plan is to build a robot that uses its cameras, sensors and machine learning software to see the world around it. It will then use this information to learn and perform tasks, without needing to be coded for every potential future situation.
X, formally known as Google X, is a subsidiary of Alphabet (also owns Google). The company’s mission is to develop “moonshot” technologies that make the world a “radically better” place. Their previous projects include Waymo, Google’s self-driving car, Wing and Malta.
The Everyday Robot project is building a new type of learning robot—one that can eventually learn to help everyone, every day.
A day in the life of an Everyday Robot
Everyday robots are training at X’s own office across Northern California. During the day, the robots sort garbage into different piles. For instance, they go through the trash and create piles for compost, landfill, and recycling. In the gif below, the robot is separating tin cans and styrofoam cups.
At night, virtual robots practice grabbing things in simulated workplaces. Both, real and virtual data, are then combined together, and given to the robots in a system update every two weeks. According to X, constant practice is making the robots good at sorting. They only sort 5% of the trash in the wrong place, whereas X employees put 20% of the trash in the wrong bin.
In fact, sorting trash is a test from X for the robots. The plan is to create robots that learn from experience. Traditionally, robots follow very specific instructions that have been coded into their system. These specific sets of instructions work well in controlled environments, like a factory. Where the robot performs the same task daily, and its environment doesn’t change. However, a robot that is helping elderly people in old homes, comes across situations that change daily. For instance, one day the chair is beside the bed and the next day in front. It is hard for coders to predict every situation and write code that responds to all of those circumstances.
You May Like: Bluetooth Devices Vulnerable to Hacking
Conclusion
In short, I think this project has great potential for the future. However, X still has a long way to go before the robots can truly become useful and start getting used in homes.
Do you think Everyday Robot is a useful project? Will X be able to produce fully functional robots that learn from the environment around it? Let us know your thoughts in the comments below!
Continue to check out Maticstoday for the latest news items, product reviews, security practices, and video game discussions.
Source: X, the moonshot factory
