Parallel Multiscale Autoregressive Density Estimation

Scott Reed, Aäron Oord, Nal Kalchbrenner, Sergio Gómez Colmenarejo, Ziyu Wang, Yutian Chen, Dan Belov, Nando Freitas
Proceedings of the 34th International Conference on Machine Learning, PMLR 70:2912-2921, 2017.

Abstract

PixelCNN achieves state-of-the-art results in density estimation for natural images. Although training is fast, inference is costly, requiring one network evaluation per pixel; O(N) for N pixels. This can be sped up by caching activations, but still involves generating each pixel sequentially. In this work, we propose a parallelized PixelCNN that allows more efficient inference by modeling certain pixel groups as conditionally independent. Our new PixelCNN model achieves competitive density estimation and orders of magnitude speedup – O(log N) sampling instead of O(N) – enabling the practical generation of 512x512 images. We evaluate the model on class-conditional image generation, text-to-image synthesis, and action-conditional video generation, showing that our model achieves the best results among non-pixel-autoregressive density models that allow efficient sampling.

Cite this Paper

BibTeX
@InProceedings{pmlr-v70-reed17a, title = {Parallel Multiscale Autoregressive Density Estimation}, author = {Scott Reed and A{\"a}ron van den Oord and Nal Kalchbrenner and Sergio G{\'o}mez Colmenarejo and Ziyu Wang and Yutian Chen and Dan Belov and Nando de Freitas}, booktitle = {Proceedings of the 34th International Conference on Machine Learning}, pages = {2912--2921}, year = {2017}, editor = {Precup, Doina and Teh, Yee Whye}, volume = {70}, series = {Proceedings of Machine Learning Research}, month = {06--11 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v70/reed17a/reed17a.pdf}, url = {https://proceedings.mlr.press/v70/reed17a.html}, abstract = {PixelCNN achieves state-of-the-art results in density estimation for natural images. Although training is fast, inference is costly, requiring one network evaluation per pixel; O(N) for N pixels. This can be sped up by caching activations, but still involves generating each pixel sequentially. In this work, we propose a parallelized PixelCNN that allows more efficient inference by modeling certain pixel groups as conditionally independent. Our new PixelCNN model achieves competitive density estimation and orders of magnitude speedup – O(log N) sampling instead of O(N) – enabling the practical generation of 512x512 images. We evaluate the model on class-conditional image generation, text-to-image synthesis, and action-conditional video generation, showing that our model achieves the best results among non-pixel-autoregressive density models that allow efficient sampling.} }
Endnote
%0 Conference Paper %T Parallel Multiscale Autoregressive Density Estimation %A Scott Reed %A Aäron Oord %A Nal Kalchbrenner %A Sergio Gómez Colmenarejo %A Ziyu Wang %A Yutian Chen %A Dan Belov %A Nando Freitas %B Proceedings of the 34th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2017 %E Doina Precup %E Yee Whye Teh %F pmlr-v70-reed17a %I PMLR %P 2912--2921 %U https://proceedings.mlr.press/v70/reed17a.html %V 70 %X PixelCNN achieves state-of-the-art results in density estimation for natural images. Although training is fast, inference is costly, requiring one network evaluation per pixel; O(N) for N pixels. This can be sped up by caching activations, but still involves generating each pixel sequentially. In this work, we propose a parallelized PixelCNN that allows more efficient inference by modeling certain pixel groups as conditionally independent. Our new PixelCNN model achieves competitive density estimation and orders of magnitude speedup – O(log N) sampling instead of O(N) – enabling the practical generation of 512x512 images. We evaluate the model on class-conditional image generation, text-to-image synthesis, and action-conditional video generation, showing that our model achieves the best results among non-pixel-autoregressive density models that allow efficient sampling.
APA
Reed, S., Oord, A., Kalchbrenner, N., Colmenarejo, S.G., Wang, Z., Chen, Y., Belov, D. & Freitas, N.. (2017). Parallel Multiscale Autoregressive Density Estimation. Proceedings of the 34th International Conference on Machine Learning, in Proceedings of Machine Learning Research 70:2912-2921 Available from https://proceedings.mlr.press/v70/reed17a.html.

Related Material