A range of methods with suitable inductive biases exist to learn interpretable object-centric representations of images without supervision. However, these are largely restricted to visually simple images; robust object discovery in real-world sensory datasets remains elusive. To increase the understanding of such inductive biases, we empirically investigate the role of "reconstruction bottlenecks" for scene decomposition in GENESIS, a recent VAE-based model. We show such bottlenecks determine reconstruction and segmentation quality and critically influence model behaviour.
@article{arxiv.2007.06245,
title = {Reconstruction Bottlenecks in Object-Centric Generative Models},
author = {Martin Engelcke and Oiwi Parker Jones and Ingmar Posner},
journal= {arXiv preprint arXiv:2007.06245},
year = {2020}
}
Comments
10 pages, 7 Figures, Workshop on Object-Oriented Learning at ICML 2020