Revisit Overconfidence for OOD Detection: Reassigned Contrastive Learning with Adaptive Class-dependent Threshold
Yanan Wu, Keqing He, Yuanmeng Yan, QiXiang Gao, Zhiyuan Zeng, Fujia Zheng, Lulu Zhao, Huixing Jiang, Wei Wu, Weiran Xu
Abstract
Detecting Out-of-Domain (OOD) or unknown intents from user queries is essential in a task-oriented dialog system. A key challenge of OOD detection is the overconfidence of neural models. In this paper, we comprehensively analyze overconfidence and classify it into two perspectives: over-confident OOD and in-domain (IND). Then according to intrinsic reasons, we respectively propose a novel reassigned contrastive learning (RCL) to discriminate IND intents for over-confident OOD and an adaptive class-dependent local threshold mechanism to separate similar IND and OOD intents for over-confident IND. Experiments and analyses show the effectiveness of our proposed method for both aspects of overconfidence issues.- Anthology ID:
- 2022.naacl-main.307
- Volume:
- Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
- Month:
- July
- Year:
- 2022
- Address:
- Seattle, United States
- Venue:
- NAACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 4165–4179
- Language:
- URL:
- https://aclanthology.org/2022.naacl-main.307
- DOI:
- 10.18653/v1/2022.naacl-main.307
- Cite (ACL):
- Yanan Wu, Keqing He, Yuanmeng Yan, QiXiang Gao, Zhiyuan Zeng, Fujia Zheng, Lulu Zhao, Huixing Jiang, Wei Wu, and Weiran Xu. 2022. Revisit Overconfidence for OOD Detection: Reassigned Contrastive Learning with Adaptive Class-dependent Threshold. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 4165–4179, Seattle, United States. Association for Computational Linguistics.
- Cite (Informal):
- Revisit Overconfidence for OOD Detection: Reassigned Contrastive Learning with Adaptive Class-dependent Threshold (Wu et al., NAACL 2022)
- PDF:
- https://preview.aclanthology.org/ingestion-script-update/2022.naacl-main.307.pdf
- Code
- pris-nlp/naacl2022-reassigned_contrastive_learning_ood
- Data
- SNIPS