Abstract
We simulate first- and second-order context overlap and show that Skip-Gram with Negative Sampling is similar to Singular Value Decomposition in capturing second-order co-occurrence information, while Pointwise Mutual Information is agnostic to it. We support the results with an empirical study finding that the models react differently when provided with additional second-order information. Our findings reveal a basic property of Skip-Gram with Negative Sampling and point towards an explanation of its success on a variety of tasks.- Anthology ID:
- W19-4803
- Volume:
- Proceedings of the 2019 ACL Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP
- Month:
- August
- Year:
- 2019
- Address:
- Florence, Italy
- Editors:
- Tal Linzen, Grzegorz Chrupała, Yonatan Belinkov, Dieuwke Hupkes
- Venue:
- BlackboxNLP
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 24–30
- Language:
- URL:
- https://aclanthology.org/W19-4803
- DOI:
- 10.18653/v1/W19-4803
- Cite (ACL):
- Dominik Schlechtweg, Cennet Oguz, and Sabine Schulte im Walde. 2019. Second-order Co-occurrence Sensitivity of Skip-Gram with Negative Sampling. In Proceedings of the 2019 ACL Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP, pages 24–30, Florence, Italy. Association for Computational Linguistics.
- Cite (Informal):
- Second-order Co-occurrence Sensitivity of Skip-Gram with Negative Sampling (Schlechtweg et al., BlackboxNLP 2019)
- PDF:
- https://preview.aclanthology.org/proper-vol2-ingestion/W19-4803.pdf
- Code
- Garrafao/SecondOrder