An AI robot that learns to walk, without data sets and/or visual examples: that’s Albert. And you can see how it works in this fun video.
In the video, which has been on YouTube for two weeks, you can see Albert. Albert is an AI robot that learns to walk. He does this by coming from five rooms, a kind of “escape rooms”. But there is something special about this process. In most cases, an AI robot learns to walk using a set of data and/or images of someone walking. In this case, the creator of Albert, whose name I don’t even know, chose to program it from scratch.
Deep Reinforcement Learning
Albert and the “learning environment” was created thanks to Deep Reinforcement Learning, a method of machine learning in which Albert is rewarded for performing actions correctly and punished for performing them incorrectly. For each attempt Albert makes, a score is calculated for how “good” it was, and small calculated adjustments are made to his brain to encourage the behaviors that lead to a higher score.
Albert is controlled by an artificial brain (neural network) made up of five layers. The first layer includes the inputs (the information it receives before taking an action, such as the position and speed of its limbs) and the last layer tells it what actions to take. In the three middle layers, calculations are performed to convert inputs into actions. It is therefore updated after each attempt by Albert.
Come on, AI Albert, come on!
This is how Albert eventually learns to walk, overcome obstacles and challenges, and turn. Be sure to watch the video, as it is a fun and special process to watch!
“Devoted bacon guru. Award-winning explorer. Internet junkie. Web lover.”