Herding Dynamical Weights to Learn

Abstract

A new ``herding'' algorithm is proposed which directly converts observed moments into a sequence of pseudo-samples. The pseudo-samples respect the moment constraints and may be used to estimate (unobserved) quantities of interest. The procedure allows us to sidestep the usual approach of first learning a joint model (which is intractable) and then sampling from that model (which can easily get stuck in a local mode). Moreover, the algorithm is fully deterministic, avoiding random number generation) and does not need expensive operations such as exponentiation.

Cite

Text

Welling. "Herding Dynamical Weights to Learn." International Conference on Machine Learning, 2009. doi:10.1145/1553374.1553517

Markdown

[Welling. "Herding Dynamical Weights to Learn." International Conference on Machine Learning, 2009.](https://mlanthology.org/icml/2009/welling2009icml-herding/) doi:10.1145/1553374.1553517

BibTeX

@inproceedings{welling2009icml-herding,
  title     = {{Herding Dynamical Weights to Learn}},
  author    = {Welling, Max},
  booktitle = {International Conference on Machine Learning},
  year      = {2009},
  pages     = {1121-1128},
  doi       = {10.1145/1553374.1553517},
  url       = {https://mlanthology.org/icml/2009/welling2009icml-herding/}
}