PixelTransformer: Sample Conditioned Signal Generation
Abstract
We propose a generative model that can infer a distribution for the underlying spatial signal conditioned on sparse samples e.g. plausible images given a few observed pixels. In contrast to sequential autoregressive generative models, our model allows conditioning on arbitrary samples and can answer distributional queries for any location. We empirically validate our approach across three image datasets and show that we learn to generate diverse and meaningful samples, with the distribution variance reducing given more observed pixels. We also show that our approach is applicable beyond images and can allow generating other types of spatial outputs e.g. polynomials, 3D shapes, and videos.
Cite
Text
Tulsiani and Gupta. "PixelTransformer: Sample Conditioned Signal Generation." International Conference on Machine Learning, 2021.Markdown
[Tulsiani and Gupta. "PixelTransformer: Sample Conditioned Signal Generation." International Conference on Machine Learning, 2021.](https://mlanthology.org/icml/2021/tulsiani2021icml-pixeltransformer/)BibTeX
@inproceedings{tulsiani2021icml-pixeltransformer,
title = {{PixelTransformer: Sample Conditioned Signal Generation}},
author = {Tulsiani, Shubham and Gupta, Abhinav},
booktitle = {International Conference on Machine Learning},
year = {2021},
pages = {10455-10464},
volume = {139},
url = {https://mlanthology.org/icml/2021/tulsiani2021icml-pixeltransformer/}
}