Shewmake, C., Buracas, D., Lillemark, H., Shin, J., Bekkers, E., Miolane, N., Olshausen, B.
Abstract
We propose a hierarchical neural network architecture for unsupervised learning of equiv-ariant part-whole decompositions of visual scenes. In contrast to the global equivariance of group-equivariant networks, the proposed architecture exhibits equivariance to part-whole transformations throughout the hierarchy, which we term hierarchical equivariance. The model achieves these structured internal representations via hierarchical Bayesian inference, which gives rise to rich bottom-up, top-down, and lateral information flows, hypothesized to underlie the mechanisms of perceptual inference in visual cortex. We demonstrate these useful properties of the model on a simple dataset of scenes with multiple objects under independent rotations and translations.
Citation
Shewmake, C., Buracas, D., Lillemark, H., Shin, J., Bekkers, E., Miolane, N., & Olshausen, B. (2023, December). Visual Scene Representation with Hierarchical Equivariant Sparse Coding. Proceedings of Machine Learning Research-NeurIPS workshop.
BibTeX
@misc{shewmake2023visual,
title={Visual Scene Representation with Hierarchical Equivariant Sparse Coding},
author={Shewmake and Buracas and Lillemarkn and Shin and Bekkers and Miolane and Olshausen},
year={2023},
organization={Proceedings of Machine Learning Research-NeurIPS workshop}
}
