Abstract
Online discussions are abundant with opinions towards a common topic, and identifying (dis)agreement between a pair of comments enables many opinion mining applications. Realizing the increasing needs to analyze opinions for emergent new topics that however tend to lack annotations, we present the first meta-learning approach for few-shot (dis)agreement identification that can be quickly applied to analyze opinions for new topics with few labeled instances. Furthermore, we enhance the meta-learner’s domain generalization ability from two perspectives. The first is domain-invariant regularization, where we design a lexicon-based regularization loss to enable the meta-learner to learn domain-invariant cues. The second is domain-aware augmentation, where we propose domain-aware task augmentation for meta-training to learn domain-specific expressions. In addition to using an existing dataset, we also evaluate our approach on two very recent new topics, mask mandate and COVID vaccine, using our newly annotated datasets containing 1.5k and 1.4k SubReddits comment pairs respectively. Extensive experiments on three domains/topics demonstrate the effectiveness of our meta-learning approach.- Anthology ID:
- 2022.findings-emnlp.409
- Volume:
- Findings of the Association for Computational Linguistics: EMNLP 2022
- Month:
- December
- Year:
- 2022
- Address:
- Abu Dhabi, United Arab Emirates
- Venue:
- Findings
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 5581–5593
- Language:
- URL:
- https://aclanthology.org/2022.findings-emnlp.409
- DOI:
- Cite (ACL):
- Yuanyuan Lei and Ruihong Huang. 2022. Few-Shot (Dis)Agreement Identification in Online Discussions with Regularized and Augmented Meta-Learning. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 5581–5593, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
- Cite (Informal):
- Few-Shot (Dis)Agreement Identification in Online Discussions with Regularized and Augmented Meta-Learning (Lei & Huang, Findings 2022)
- PDF:
- https://preview.aclanthology.org/ingestion-script-update/2022.findings-emnlp.409.pdf