Prototypical Verbalizer for Prompt-based Few-shot Tuning
This work addresses the problem of reducing human effort and domain knowledge dependency in few-shot learning for NLP researchers and practitioners, offering an incremental improvement over existing automatic verbalizer methods.
The paper tackles the challenge of automatically constructing verbalizers for prompt-based few-shot tuning by proposing ProtoVerb, which learns prototype vectors from training data via contrastive learning, achieving significant performance improvements over existing automatic verbalizers, especially with extremely scarce data, and even boosting untuned pre-trained language models.
Prompt-based tuning for pre-trained language models (PLMs) has shown its effectiveness in few-shot learning. Typically, prompt-based tuning wraps the input text into a cloze question. To make predictions, the model maps the output words to labels via a verbalizer, which is either manually designed or automatically built. However, manual verbalizers heavily depend on domain-specific prior knowledge and human efforts, while finding appropriate label words automatically still remains challenging.In this work, we propose the prototypical verbalizer (ProtoVerb) which is built directly from training data. Specifically, ProtoVerb learns prototype vectors as verbalizers by contrastive learning. In this way, the prototypes summarize training instances and are able to enclose rich class-level semantics. We conduct experiments on both topic classification and entity typing tasks, and the results demonstrate that ProtoVerb significantly outperforms current automatic verbalizers, especially when training data is extremely scarce. More surprisingly, ProtoVerb consistently boosts prompt-based tuning even on untuned PLMs, indicating an elegant non-tuning way to utilize PLMs. Our codes are avaliable at https://github.com/thunlp/OpenPrompt.