Clip-Tuning: Towards Derivative-free Prompt Learning with a Mixture of Rewards
Yekun Chai, Shuohuan Wang, Yu Sun, Hao Tian, Hua Wu, Haifeng Wang
Abstract
Derivative-free prompt learning has emerged as a lightweight alternative to prompt tuning, which only requires model inference to optimize the prompts. However, existing work did not take full advantage of the over-parameterized characteristics of large pre-trained language models (PLMs). In this paper, we propose Clip-Tuning, a simple yet effective method that adopts diverse frozen “thinned” networks of PLMs to obtain *a mixture of rewards* and thus advance the derivative-free prompt learning. The thinned networks consist of all the hidden units that survive a stationary dropout strategy, whose inference predictions reflect an ensemble of partial views over prompted training samples. Our method outperforms previous gradient-free prompt learning methods and achieves parity with gradient-based counterparts on seven language understanding benchmarks under few-shot settings.- Anthology ID:
- 2022.findings-emnlp.8
- 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:
- 108–117
- Language:
- URL:
- https://aclanthology.org/2022.findings-emnlp.8
- DOI:
- Cite (ACL):
- Yekun Chai, Shuohuan Wang, Yu Sun, Hao Tian, Hua Wu, and Haifeng Wang. 2022. Clip-Tuning: Towards Derivative-free Prompt Learning with a Mixture of Rewards. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 108–117, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
- Cite (Informal):
- Clip-Tuning: Towards Derivative-free Prompt Learning with a Mixture of Rewards (Chai et al., Findings 2022)
- PDF:
- https://preview.aclanthology.org/ingestion-script-update/2022.findings-emnlp.8.pdf