Adaptive Prompt Learning-Based Few-Shot Sentiment Analysis

In the field of natural language processing, sentiment analysis via deep learning has a excellent performance by using large labeled datasets. Meanwhile, labeled data are insufficient in many sentiment analysis tasks, and obtaining these data is time-consuming and laborious. Prompt learning devotes to resolving the data deficiency by reformulating downstream tasks with the help of prompt. The model performance of this method depends on the quality of the prompt. This paper proposes an adaptive prompting (AP) construction strategy using seq2seq-attention structure to acquire the semantic information of the input sequence. Our method of dynamically constructing adaptive prompts can not only improve the quality of prompt, but also can effectively generalize to other fields by constructing a pre-trained prompt with existing public labeled data. The experimental results on FewCLUE datasets demonstrate that the proposed method AP can effectively construct appropriate adaptive prompt regardless of the quality of hand-crafted prompt and outperform the state-of-the-art baselines.