Covariant, which is working with Knapp, built software that could learn through trial and error. First, the system learned from a digital simulation of the task — a virtual recreation of a bin filled with random items. Then, when Mr. Chen and his colleagues transferred this software to a robot, it could pick up items in the real world.A pit boss does the same except the pit boss incentivizes the traders to conform to the marker making error allowed. There is no such feed back in this warehouse system, but with additional learning the solution can be found, within some bounded error.
The robot could continue to learn as it sorted through items it had never seen before. Inside the German warehouse, the robot can pick and sort more than 10,000 different items, and it does this with more than 99 percent accuracy, according to Covariant.
The math is similar to our pit bosses. Organize the flow into a 'value added net' that produces separable features in some complete sequence. Thus the semantic decision tree can always be kept minimal as long as the bot continually checks the feature sets for redundancy.
It is still an entropy maximizing process, a virtual loop closing process in the structured queues. Under the hood we will find most of the auto-learning bots to be special purpose pit bosses. This is a co-invention, sandbox gets partial credit.
No comments:
Post a Comment