Researchers add ‘common sense’ to computers

Researchers from UC San Diego and UCLA have used a Google Labs widget to make computers behave in a more ‘common sense’ way.

The scientists have given a computer the ability to use context to help identify objects in photographs in an automated image labelling system.

A conventional automated object identifier might scan an image and identify ‘person’, ‘tennis racket’, ‘tennis court’ and ‘lemon’.

The researchers have now been able to add a context check, which would re-label ‘lemon’ as ‘tennis ball’.

This ability is provided via Google Sets, which creates lists of related items or objects from just a few examples.

The scientists have published a paper showing how this common sense can provide contextual information to improve the accuracy of automated image labelling systems.

The Objects in Context paper (PDF) was presented at the 11th IEEE International Conference on Computer Vision in Rio de Janeiro.