Generalized Earley Parser: Bridging Symbolic Grammars and Sequence Data for Future Prediction

Siyuan Qi, Baoxiong Jia, Song-Chun Zhu
Proceedings of the 35th International Conference on Machine Learning, PMLR 80:4171-4179, 2018.

Abstract

Future predictions on sequence data (e.g., videos or audios) require the algorithms to capture non-Markovian and compositional properties of high-level semantics. Context-free grammars are natural choices to capture such properties, but traditional grammar parsers (e.g., Earley parser) only take symbolic sentences as inputs. In this paper, we generalize the Earley parser to parse sequence data which is neither segmented nor labeled. This generalized Earley parser integrates a grammar parser with a classifier to find the optimal segmentation and labels, and makes top-down future predictions. Experiments show that our method significantly outperforms other approaches for future human activity prediction.

Cite this Paper

BibTeX
@InProceedings{pmlr-v80-qi18a, title = {Generalized Earley Parser: Bridging Symbolic Grammars and Sequence Data for Future Prediction}, author = {Qi, Siyuan and Jia, Baoxiong and Zhu, Song-Chun}, booktitle = {Proceedings of the 35th International Conference on Machine Learning}, pages = {4171--4179}, year = {2018}, editor = {Dy, Jennifer and Krause, Andreas}, volume = {80}, series = {Proceedings of Machine Learning Research}, month = {10--15 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v80/qi18a/qi18a.pdf}, url = {https://proceedings.mlr.press/v80/qi18a.html}, abstract = {Future predictions on sequence data (e.g., videos or audios) require the algorithms to capture non-Markovian and compositional properties of high-level semantics. Context-free grammars are natural choices to capture such properties, but traditional grammar parsers (e.g., Earley parser) only take symbolic sentences as inputs. In this paper, we generalize the Earley parser to parse sequence data which is neither segmented nor labeled. This generalized Earley parser integrates a grammar parser with a classifier to find the optimal segmentation and labels, and makes top-down future predictions. Experiments show that our method significantly outperforms other approaches for future human activity prediction.} }
Endnote
%0 Conference Paper %T Generalized Earley Parser: Bridging Symbolic Grammars and Sequence Data for Future Prediction %A Siyuan Qi %A Baoxiong Jia %A Song-Chun Zhu %B Proceedings of the 35th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2018 %E Jennifer Dy %E Andreas Krause %F pmlr-v80-qi18a %I PMLR %P 4171--4179 %U https://proceedings.mlr.press/v80/qi18a.html %V 80 %X Future predictions on sequence data (e.g., videos or audios) require the algorithms to capture non-Markovian and compositional properties of high-level semantics. Context-free grammars are natural choices to capture such properties, but traditional grammar parsers (e.g., Earley parser) only take symbolic sentences as inputs. In this paper, we generalize the Earley parser to parse sequence data which is neither segmented nor labeled. This generalized Earley parser integrates a grammar parser with a classifier to find the optimal segmentation and labels, and makes top-down future predictions. Experiments show that our method significantly outperforms other approaches for future human activity prediction.
APA
Qi, S., Jia, B. & Zhu, S.. (2018). Generalized Earley Parser: Bridging Symbolic Grammars and Sequence Data for Future Prediction. Proceedings of the 35th International Conference on Machine Learning, in Proceedings of Machine Learning Research 80:4171-4179 Available from https://proceedings.mlr.press/v80/qi18a.html.

Related Material