Researchers from Google's DeepMind division say they have shown how their differentiable neural computer is able to process information using so-called 'slow' thinking thought processes.
The researchers published their results in a paper in Nature explaining how their differentiable neural computer, which is an algorithm and not a physical computer, emulates human slow thinking by processing often disparate information to solve problems.
It does this with a controller that combines recorded information with neural networks. Like a slow thinking human thought process, it analyses stored memories and a timeline when each memory was recorded until it successfully finds a solution to a problem.
In their paper, the DeepMind researchers showed how they were able to calculate routes on the London Underground using the differential neural computer. They reported they will apply the process to tasks involving much larger data pools in the future.