Visual Analogy Extrapolation Challenge (VAEC)

Webb, Taylor; Dulberg, Zachary; Frankland, Steven; Petrov, Alexander; O'Reilly, Randall; Cohen, Jonathan
Issue date: 2020
Rights:
Creative Commons Attribution 4.0 International (CC BY)
Cite as:
Webb, Taylor, Dulberg, Zachary, Frankland, Steven, Petrov, Alexander, O'Reilly, Randall, & Cohen, Jonathan. (2020). Visual Analogy Extrapolation Challenge (VAEC) [Data set]. Princeton University. https://doi.org/10.34770/81bg-rt16
@electronic{webb_taylor_2020,
  author      = {Webb, Taylor and
                Dulberg, Zachary and
                Frankland, Steven and
                Petrov, Alexander and
                O'Reilly, Randall and
                Cohen, Jonathan},
  title       = {{Visual Analogy Extrapolation Challenge (
                VAEC)}},
  publisher   = {{Princeton University}},
  year        = 2020,
  url         = {https://doi.org/10.34770/81bg-rt16}
}
Description:

Extrapolation -- the ability to make inferences that go beyond the scope of one's experiences -- is a hallmark of human intelligence. By contrast, the generalization exhibited by contemporary neural network algorithms is largely limited to interpolation between data points in their training corpora. In this paper, we consider the challenge of learning representations that support extrapolation. We introduce a novel visual analogy benchmark that allows the graded evaluation of extrapolation as a function of distance from the convex domain defined by the training data. We also introduce a simple technique, context normalization, that encourages representations that emphasize the relations between objects. We find that this technique enables a significant improvement in the ability to extrapolate, considerably outperforming a number of competitive techniques.

Show More