POUF: Prompt-oriented unsupervised fine-tuning for large pre-trained models
This addresses the critical limitation of requiring labeled data for fine-tuning in NLP and vision tasks, offering a practical solution for scenarios with limited annotations.
The paper tackles the problem of adapting large pre-trained models to downstream tasks without labeled data by proposing an unsupervised fine-tuning framework that aligns discrete distributions from prompts and target data, achieving consistent improvements across 28 tasks in image classification, sentiment analysis, and natural language inference.
Through prompting, large-scale pre-trained models have become more expressive and powerful, gaining significant attention in recent years. Though these big models have zero-shot capabilities, in general, labeled data are still required to adapt them to downstream tasks. To overcome this critical limitation, we propose an unsupervised fine-tuning framework to directly fine-tune the model or prompt on the unlabeled target data. We demonstrate how to apply our method to both language-augmented vision and masked-language models by aligning the discrete distributions extracted from the prompts and target data. To verify our approach's applicability, we conduct extensive experiments on image classification, sentiment analysis, and natural language inference tasks. Across 13 image-related tasks and 15 language-related ones, the proposed approach achieves consistent improvements over the baselines.