The road to the rise of robots

According to Gigaom editor and Derrick Harris, a group called RoboEarth, at the Federal Institute of Technology in Zurich, Switzerland, is currently releasing the first platform version of cloud computing for robots. They compare the database, so that the robot can be in different parts of the world, in different situations to their own task update processing, in order to achieve the best results.

Briefly described below:

A group of European researchers released the first version of the cloud computing platform that robots use to help them gain access to powerful virtual resources. In essence, they think of robots as any other device, such as desktops, tablets, or mobile phones--running web applications, allowing robots to learn from each other, and doing more than just updating the screen display.

A team called RoboEarth at the Federal Institute of Technology in Zurich, Switzerland, conducted a study of the project, the key of which is the cloud software platform known as Rapyuta. It works in a very simple way: the robot communicates with a cloud-based application platform, performs computational tasks, and connects to the full information of a cloud database, such as maps, pictures, languages, and other Web services. The robot itself is equivalent to many hardware terminals, equipped with sensors and moving parts, but is limited by processing power and data storage.

Of course, we've been looking for ideas to connect to our mainframe computers through cloud computing, web and mobile apps. Why do you have to confine these things to the body of a robot in the context of today's computing power, memory, storage, and data that can be extended indefinitely over the Internet? With a powerful network connector inside the robot, it does not require a large hard drive or the latest, best processor to do much work.

However, one thing RoboEarth do is a little different, they through the database, so that robots can be in different parts of the world in different situations for their own task updates. This is actually machine learning, allowing the robot to learn from the experience of other robots. This makes robots more useful to people who do not program them, and may enable robots to perform more and more varied tasks.

However, it is speculated that some back-end statistical machine learning can make robots smarter, they produce more and more data and initial prototype of the model.

This project sounds a bit like the symbol of Skynet (Skynet) and the rise of the machine in Terminator, but it's a bit exaggerated. After all, the machines themselves do not communicate with each other, but through human operations a centralized computing infrastructure. It is similar to IBM's Watson system and is good at answering questions, but its excellent answer questions are built on the same excellent information database and algorithms.