Automatically Identifying Words That Can Serve as Labels for Few-Shot Text Classification
This addresses the need for domain expertise in few-shot learning, though it is incremental as it builds on existing cloze-based methods.
The paper tackles the problem of manually defining label-to-word mappings for few-shot text classification by devising an approach that automatically finds such mappings from small training data, achieving performance almost as good as hand-crafted mappings for several tasks.
A recent approach for few-shot text classification is to convert textual inputs to cloze questions that contain some form of task description, process them with a pretrained language model and map the predicted words to labels. Manually defining this mapping between words and labels requires both domain expertise and an understanding of the language model's abilities. To mitigate this issue, we devise an approach that automatically finds such a mapping given small amounts of training data. For a number of tasks, the mapping found by our approach performs almost as well as hand-crafted label-to-word mappings.