Attentive Group Equivariant Convolutional Networks
This work addresses a problem in machine learning for computer vision by improving the performance and interpretability of group equivariant networks, though it is incremental as it builds on prior attention and group convolution methods.
The paper tackles the limitation of group convolutional networks in learning meaningful relationships among symmetry patterns by introducing attentive group equivariant convolutions, which apply attention during convolution to highlight plausible symmetry combinations, and shows that this approach consistently outperforms conventional group convolutional networks on benchmark image datasets.
Although group convolutional networks are able to learn powerful representations based on symmetry patterns, they lack explicit means to learn meaningful relationships among them (e.g., relative positions and poses). In this paper, we present attentive group equivariant convolutions, a generalization of the group convolution, in which attention is applied during the course of convolution to accentuate meaningful symmetry combinations and suppress non-plausible, misleading ones. We indicate that prior work on visual attention can be described as special cases of our proposed framework and show empirically that our attentive group equivariant convolutional networks consistently outperform conventional group convolutional networks on benchmark image datasets. Simultaneously, we provide interpretability to the learned concepts through the visualization of equivariant attention maps.