Human-level concept learning through probabilistic program induction -- visual Turing tests

By Brenden Lake, Ruslan Salakhutdinov, and Joshua Tenenbaum
Email: brenden at-sign nyu dot edu

Try our "visual Turing tests" to see if you can tell the difference between human and machine behavior. The code and images for running these experiments are available on github.

Four tasks for comparing humans versus machines

Generating new exemplars

Generating new exemplars (dynamic)

Generating new concepts (from type)

Generating new concepts (unconstrained)

Additional options

You can a different version with complete feedback here. You can also try the visual Turing tests for the lesioned models by following this link.