We consider a class of visual analogical reasoning problems that involve discovering the sequence of transformations by which pairs of input/output images are related, so as to analogously transform future inputs. This program synthesis task can be easily solved via symbolic search. Using a variation of the `neural analogical reasoning' approach of (Velickovic and Blundell 2021), we instead search for a sequence of elementary neural network transformations that manipulate distributed representations derived from a symbolic space, to which input images are directly encoded. We evaluate the extent to which our `neural reasoning' approach generalizes for images with unseen shapes and positions.
@article{arxiv.2111.10361,
title = {Solving Visual Analogies Using Neural Algorithmic Reasoning},
author = {Atharv Sonwane and Gautam Shroff and Lovekesh Vig and Ashwin Srinivasan and Tirtharaj Dash},
journal= {arXiv preprint arXiv:2111.10361},
year = {2021}
}
Comments
20 pages. Contains extended abstract accepted at the AAAI-22 Student Abstract and Poster Program along with relevent supplementary material