Few-shot Unified Question Answering: Tuning Models or Prompts?

Question-answering (QA) tasks often investigate specific question types, knowledge domains, or reasoning skills, leading to specialized models catering to specific categories of QA tasks. While recent research has explored the idea of unified QA models, such models are usually explored for high-resource scenarios and require re-training to extend their capabilities. To overcome these drawbacks, the paper explores the potential of two paradigms of model tuning and prompt tuning for unified QA under a low-resource setting. The paper provides an exhaustive analysis of their applicability using 16 QA datasets spanning across various domains, revealing that prompt tuning outperforms model tuning in a fewshot setting with good initialization for out-ofdistribution target tasks. The study also shows that parameter-sharing results in superior fewshot performance, and simple knowledge transfer techniques for prompt initialization can be effective in a low-resource regime. The research offers insights into the advantages and limitations of prompt tuning for unified QA in a few-shot setting, contributing to the development of effective and efficient systems in low-resource scenarios.