Spatial reasoning arises from combining relational networks and convoluted neural networks

Without knowing how to put it into words, I was recently thinking about this ability in artificial intelligence. Now I know it's called "spatial reasoning." Machine intelligence is moving ahead much faster than I realized before I started looking closely at it in the past month.

humans possess something like an “intuitive physics engine,” an algorithm for extrapolating three-dimensionality from flat images and comparing objects within it to other objects. This kind of spatial reasoning has proved difficult for computers, at least until now. Using a combination of relational networks and convoluted neural networks, the DeepMind system can answer questions concerning the relation of objects within an image. (from https://www.extremetech.com/extreme/251126-ai-acquires-spatial-reasoning-abilities-another-victory-machine-overlords)

Add a comment

HTML code is displayed as text and web addresses are automatically converted.

Page top