Unsupervised Search-Based Structured Prediction
Abstract
We describe an adaptation and application of a search-based structured prediction algorithm "Searn" to unsupervised learning problems. We show that it is possible to reduce unsupervised learning to supervised learning and demonstrate a high-quality un-supervised shift-reduce parsing model. We additionally show a close connection between unsupervised Searn and expectation maximization. Finally, we demonstrate the efficacy of a semi-supervised extension. The key idea that enables this is an application of the predict-self idea for unsupervised learning.
Cite
Text
Iii. "Unsupervised Search-Based Structured Prediction." International Conference on Machine Learning, 2009. doi:10.1145/1553374.1553401Markdown
[Iii. "Unsupervised Search-Based Structured Prediction." International Conference on Machine Learning, 2009.](https://mlanthology.org/icml/2009/iii2009icml-unsupervised/) doi:10.1145/1553374.1553401BibTeX
@inproceedings{iii2009icml-unsupervised,
title = {{Unsupervised Search-Based Structured Prediction}},
author = {Iii, Hal Daumé},
booktitle = {International Conference on Machine Learning},
year = {2009},
pages = {209-216},
doi = {10.1145/1553374.1553401},
url = {https://mlanthology.org/icml/2009/iii2009icml-unsupervised/}
}