CLAY: Learning to Denoise Raw Mobile UI Layouts for Improving Datasets at Scale

This repository will contain the model code and the experimental framework for "Learning to Denoise Raw Mobile UI Layouts for Improving Datasets at Scale" by Gang Li, Gilles Baechler, Manuel Tragut, Yang Li, which is conditionally accepted by CHI 2022.

The CLAY dataset used by the paper can be downloaded from https://github.com/google-research-datasets/clay.

The Transformer-based object typing model can be found here.

BibTex for citation:

@InProceedings{clay,
  title     = "Learning to Denoise Raw Mobile UI Layouts for Improving Datasets at Scale",
  booktitle = "Proceedings of the 2022 {CHI} Conference on Human Factors in
               Computing Systems",
  author    = "Li, Gang and Baechler, Gilles and Tragut, Manuel and Li, Yang",
  publisher = "Association for Computing Machinery",
  pages     = "1--13",
  month     =  may,
  year      =  2022,
  address   = "New Orleans, LA, USA",
  url = {https://doi.org/10.1145/3491102.3502042}
}