Selective Token Generation for Few-shot Natural Language Generation
- Authors
- Jo, Daejin; Kwon, Taehwan; Kim, Eun Sol; Kim, Sungwoong
- Issue Date
- Oct-2022
- Publisher
- Association for Computational Linguistics
- Citation
- International Conference on Computational Linguistics, pp.5837 - 5856
- Indexed
- OTHER
- Journal Title
- International Conference on Computational Linguistics
- Start Page
- 5837
- End Page
- 5856
- URI
- https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/188948
- DOI
- 10.48550/arXiv.2209.08206
- Abstract
- Natural language modeling with limited training data is a challenging problem, and manyalgorithms make use of large-scale pretrainedlanguage models (PLMs) for this due to itsgreat generalization ability. Among them, additive learning that incorporates a task-specificadapter on top of the fixed large-scale PLMhas been popularly used in the few-shot setting. However, this added adapter is still easyto disregard the knowledge of the PLM especially for few-shot natural language generation (NLG) since an entire sequence is usuallygenerated by only the newly trained adapter.Therefore, in this work, we develop a noveladditive learning algorithm based on reinforcement learning (RL) that selectively outputs language tokens between the task-general PLMand the task-specific adapter during both training and inference. This output token selection over the two generators allows the adapterto take into account solely the task-relevantparts in sequence generation, and thereforemakes it more robust to overfitting as well asmore stable in RL training. In addition, toobtain the complementary adapter from thePLM for each few-shot task, we exploit aseparate selecting module that is also simultaneously trained using RL. Experimental results on various few-shot NLG tasks including question answering, data-to-text generation and text summarization demonstrate thatthe proposed selective token generation significantly outperforms the previous additive learning algorithms based on the PLMs.
- Files in This Item
-
Go to Link
- Appears in
Collections - 서울 공과대학 > 서울 컴퓨터소프트웨어학부 > 1. Journal Articles
![qrcode](https://api.qrserver.com/v1/create-qr-code/?size=55x55&data=https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/188948)
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.