English

Dilated DenseNets for Relational Reasoning

Machine Learning 2018-11-02 v1 Machine Learning

Abstract

Despite their impressive performance in many tasks, deep neural networks often struggle at relational reasoning. This has recently been remedied with the introduction of a plug-in relational module that considers relations between pairs of objects. Unfortunately, this is combinatorially expensive. In this extended abstract, we show that a DenseNet incorporating dilated convolutions excels at relational reasoning on the Sort-of-CLEVR dataset, allowing us to forgo this relational module and its associated expense.

Cite

@article{arxiv.1811.00410,
  title  = {Dilated DenseNets for Relational Reasoning},
  author = {Antreas Antoniou and Agnieszka Słowik and Elliot J. Crowley and Amos Storkey},
  journal= {arXiv preprint arXiv:1811.00410},
  year   = {2018}
}

Comments

Extended Abstract

R2 v1 2026-06-23T05:00:44.613Z