Robustness of Demonstration-based Learning Under Limited Data Scenario

TL;DR

This paper investigates the robustness of demonstration-based learning in language models under limited data conditions, revealing that even random demonstrations can improve performance and confidence, with factors like length and relevance affecting outcomes.

Contribution

It provides a systematic analysis of how different types of demonstrations influence model robustness and performance in few-shot learning scenarios.

Findings

Random token demonstrations still improve few-shot learning.

Demonstration length and token relevance significantly impact performance.

Demonstrations increase model confidence in superficial patterns.

Abstract

Demonstration-based learning has shown great potential in stimulating pretrained language models' ability under limited data scenario. Simply augmenting the input with some demonstrations can significantly improve performance on few-shot NER. However, why such demonstrations are beneficial for the learning process remains unclear since there is no explicit alignment between the demonstrations and the predictions. In this paper, we design pathological demonstrations by gradually removing intuitively useful information from the standard ones to take a deep dive of the robustness of demonstration-based sequence labeling and show that (1) demonstrations composed of random tokens still make the model a better few-shot learner; (2) the length of random demonstrations and the relevance of random tokens are the main factors affecting the performance; (3) demonstrations increase the confidence of model predictions on captured superficial patterns. We have publicly released our code at https://github.com/SALT-NLP/RobustDemo.