Multifactor Gaussian Process Models for Style-Content Separation
Abstract
We introduce models for density estimation with multiple, hidden, continuous factors. In particular, we propose a generalization of multilinear models using nonlinear basis functions. By marginalizing over the weights, we obtain a multifactor form of the Gaussian process latent variable model. In this model, each factor is kernelized independently, allowing nonlinear mappings from any particular factor to the data. We learn models for human locomotion data, in which each pose is generated by factors representing the person's identity, gait, and the current state of motion. We demonstrate our approach using time-series prediction, and by synthesizing novel animation from the model.
Cite
Text
Wang et al. "Multifactor Gaussian Process Models for Style-Content Separation." International Conference on Machine Learning, 2007. doi:10.1145/1273496.1273619Markdown
[Wang et al. "Multifactor Gaussian Process Models for Style-Content Separation." International Conference on Machine Learning, 2007.](https://mlanthology.org/icml/2007/wang2007icml-multifactor/) doi:10.1145/1273496.1273619BibTeX
@inproceedings{wang2007icml-multifactor,
title = {{Multifactor Gaussian Process Models for Style-Content Separation}},
author = {Wang, Jack M. and Fleet, David J. and Hertzmann, Aaron},
booktitle = {International Conference on Machine Learning},
year = {2007},
pages = {975-982},
doi = {10.1145/1273496.1273619},
url = {https://mlanthology.org/icml/2007/wang2007icml-multifactor/}
}