English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

How to improve generative modeling by better exploiting spatial regularitiesand coherence in images? We introduce a novel neural network for building imagegenerators (decoders) and apply it to variational autoencoders (VAEs). In ourspatial dependency networks (SDNs), feature maps at each level of a deep neuralnet are computed in a spatially coherent way, using a sequential gating-basedmechanism that distributes contextual information across 2-D space. We showthat augmenting the decoder of a hierarchical VAE by spatial dependency layersconsiderably improves density estimation over baseline convolutionalarchitectures and the state-of-the-art among the models within the same class.Furthermore, we demonstrate that SDN can be applied to large images bysynthesizing samples of high quality and coherence. In a vanilla VAE setting,we find that a powerful SDN decoder also improves learning disentangledrepresentations, indicating that neural architectures play an important role inthis task. Our results suggest favoring spatial dependency over convolutionallayers in various VAE settings. The accompanying source code is given athttps://github.com/djordjemila/sdn.

Spatial Dependency Networks: Neural Layers for Improved Generative Image Modeling