Shared Visual Representations of Drawing for Communication: How do different biases affect human interpretability and intent?. (arXiv:2110.08203v2 [cs.LG] UPDATED)

We present an investigation into how representational losses can affect the
drawings produced by artificial agents playing a communication game. Building
upon recent advances, we show that a combination of powerful pretrained encoder
networks, with appropriate inductive biases, can lead to agents that draw
recognisable sketches, whilst still communicating well. Further, we start to
develop an approach to help automatically analyse the semantic content being
conveyed by a sketch and demonstrate that current approaches to inducing
perceptual biases lead to …

arxiv biases communication human interpretability