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/

Global: Zendy Plus annual

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

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We address the video prediction task by putting forth a novel model thatcombines (i) our recently proposed hierarchical residual vector quantizedvariational autoencoder (HR-VQVAE), and (ii) a novel spatiotemporal PixelCNN(ST-PixelCNN). We refer to this approach as a sequential hierarchical residuallearning vector quantized variational autoencoder (S-HR-VQVAE). By leveragingthe intrinsic capabilities of HR-VQVAE at modeling still images with aparsimonious representation, combined with the ST-PixelCNN's ability athandling spatiotemporal information, S-HR-VQVAE can better deal with chiefchallenges in video prediction. These include learning spatiotemporalinformation, handling high dimensional data, combating blurry prediction, andimplicit modeling of physical characteristics. Extensive experimental resultson the KTH Human Action and Moving-MNIST tasks demonstrate that our modelcompares favorably against top video prediction techniques both in quantitativeand qualitative evaluations despite a much smaller model size. Finally, weboost S-HR-VQVAE by proposing a novel training method to jointly estimate theHR-VQVAE and ST-PixelCNN parameters.

S-HR-VQVAE: Sequential Hierarchical Residual Learning Vector Quantized Variational Autoencoder for Video Prediction