As a first step towards agents learning to communicate about their visual environment, we propose a system that, given visual representations of a referent (cat) and a context (sofa), identifies their discriminative attributes, i.e., properties that distinguish them (has_tail). Moreover, despite the lack of direct supervision at the attribute level, the model learns to assign plausible attributes to objects (sofa-has_cushion). Finally, we present a preliminary experiment confirming the referential success of the predicted discriminative attributes.
@article{arxiv.1603.02618,
title = {The red one!: On learning to refer to things based on their discriminative properties},
author = {Angeliki Lazaridou and Nghia The Pham and Marco Baroni},
journal= {arXiv preprint arXiv:1603.02618},
year = {2016}
}