Speaker Information Can Guide Models to Better Inductive Biases: A Case Study On Predicting Code-Switching
Alissa Ostapenko, Shuly Wintner, Melinda Fricke, Yulia Tsvetkov
Abstract
Natural language processing (NLP) models trained on people-generated data can be unreliable because, without any constraints, they can learn from spurious correlations that are not relevant to the task. We hypothesize that enriching models with speaker information in a controlled, educated way can guide them to pick up on relevant inductive biases. For the speaker-driven task of predicting code-switching points in English–Spanish bilingual dialogues, we show that adding sociolinguistically-grounded speaker features as prepended prompts significantly improves accuracy. We find that by adding influential phrases to the input, speaker-informed models learn useful and explainable linguistic information. To our knowledge, we are the first to incorporate speaker characteristics in a neural model for code-switching, and more generally, take a step towards developing transparent, personalized models that use speaker information in a controlled way.- Anthology ID:
- 2022.acl-long.267
- Volume:
- Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
- Month:
- May
- Year:
- 2022
- Address:
- Dublin, Ireland
- Editors:
- Smaranda Muresan, Preslav Nakov, Aline Villavicencio
- Venue:
- ACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 3853–3867
- Language:
- URL:
- https://aclanthology.org/2022.acl-long.267
- DOI:
- 10.18653/v1/2022.acl-long.267
- Cite (ACL):
- Alissa Ostapenko, Shuly Wintner, Melinda Fricke, and Yulia Tsvetkov. 2022. Speaker Information Can Guide Models to Better Inductive Biases: A Case Study On Predicting Code-Switching. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 3853–3867, Dublin, Ireland. Association for Computational Linguistics.
- Cite (Informal):
- Speaker Information Can Guide Models to Better Inductive Biases: A Case Study On Predicting Code-Switching (Ostapenko et al., ACL 2022)
- PDF:
- https://preview.aclanthology.org/improve-issue-templates/2022.acl-long.267.pdf
- Code
- ostapen/switch-and-explain