ICML 2020poster32 citations

Incremental Sampling Without Replacement for Sequence Models

Kensen Shi, David Bieber, Charles Sutton

Abstract

Sampling is a fundamental technique, and sampling without replacement is often desirable when duplicate samples are not beneficial. Within machine learning, sampling is useful for generating diverse outputs from a trained model. We present an elegant procedure for sampling without replacement from a broad class of randomized programs, including generative neural models that construct outputs sequentially. Our procedure is efficient even for exponentially-large output spaces. Unlike prior work, our approach is incremental, i.e., samples can be drawn one at a time, allowing for increased flexibility. We also present a new estimator for computing expectations from samples drawn without replacement. We show that incremental sampling without replacement is applicable to many domains, e.g., program synthesis and combinatorial optimization.

BibTeX
@InProceedings{pmlr-v119-shi20a,
  title = 	 {Incremental Sampling Without Replacement for Sequence Models},
  author =       {Shi, Kensen and Bieber, David and Sutton, Charles},
  booktitle = 	 {Proceedings of the 37th International Conference on Machine Learning},
  pages = 	 {8785--8795},
  year = 	 {2020},
  editor = 	 {III, Hal Daumé and Singh, Aarti},
  volume = 	 {119},
  series = 	 {Proceedings of Machine Learning Research},
  month = 	 {13--18 Jul},
  publisher =    {PMLR},
  pdf = 	 {http://proceedings.mlr.press/v119/shi20a/shi20a.pdf},
  url = 	 {https://proceedings.mlr.press/v119/shi20a.html},
  abstract = 	 {Sampling is a fundamental technique, and sampling without replacement is often desirable when duplicate samples are not beneficial. Within machine learning, sampling is useful for generating diverse outputs from a trained model. We present an elegant procedure for sampling without replacement from a broad class of randomized programs, including generative neural models that construct outputs sequentially. Our procedure is efficient even for exponentially-large output spaces. Unlike prior work, our approach is incremental, i.e., samples can be drawn one at a time, allowing for increased flexibility. We also present a new estimator for computing expectations from samples drawn without replacement. We show that incremental sampling without replacement is applicable to many domains, e.g., program synthesis and combinatorial optimization.}
}
Incremental Sampling Without Replacement for Sequence Models · ICML 2020