Abstract
Within the context of event modeling and understanding, we propose a new method for neural sequence modeling that takes partially-observed sequences of discrete, external knowledge into account. We construct a sequential neural variational autoencoder, which uses Gumbel-Softmax reparametrization within a carefully defined encoder, to allow for successful backpropagation during training. The core idea is to allow semi-supervised external discrete knowledge to guide, but not restrict, the variational latent parameters during training. Our experiments indicate that our approach not only outperforms multiple baselines and the state-of-the-art in narrative script induction, but also converges more quickly.- Anthology ID:
- 2021.naacl-main.374
- Volume:
- Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
- Month:
- June
- Year:
- 2021
- Address:
- Online
- Venue:
- NAACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 4701–4716
- Language:
- URL:
- https://aclanthology.org/2021.naacl-main.374
- DOI:
- 10.18653/v1/2021.naacl-main.374
- Cite (ACL):
- Mehdi Rezaee and Francis Ferraro. 2021. Event Representation with Sequential, Semi-Supervised Discrete Variables. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 4701–4716, Online. Association for Computational Linguistics.
- Cite (Informal):
- Event Representation with Sequential, Semi-Supervised Discrete Variables (Rezaee & Ferraro, NAACL 2021)
- PDF:
- https://preview.aclanthology.org/remove-xml-comments/2021.naacl-main.374.pdf