Context Variance Evaluation of Pretrained Language Models for Prompt-based Biomedical Knowledge Probing

TL;DR

This paper introduces context variance prompts and a new evaluation metric to improve the reliability of probing biomedical knowledge in pretrained language models, addressing biases and challenges in existing methods.

Contribution

It proposes a novel context variance approach and the UCM metric, enhancing the evaluation of PLMs' biomedical knowledge, especially for large-N-M and rare relations.

Findings

Context variance prompts improve robustness in knowledge probing.

UCM metric captures model understanding beyond simple recall.

Enhanced evaluation stability for large-N-M and rare relations.

Abstract

Pretrained language models (PLMs) have motivated research on what kinds of knowledge these models learn. Fill-in-the-blanks problem (e.g., cloze tests) is a natural approach for gauging such knowledge. BioLAMA generates prompts for biomedical factual knowledge triples and uses the Top-k accuracy metric to evaluate different PLMs' knowledge. However, existing research has shown that such prompt-based knowledge probing methods can only probe a lower bound of knowledge. Many factors like prompt-based probing biases make the LAMA benchmark unreliable and unstable. This problem is more prominent in BioLAMA. The severe long-tailed distribution in vocabulary and large-N-M relation make the performance gap between LAMA and BioLAMA remain notable. To address these, we introduce context variance into the prompt generation and propose a new rank-change-based evaluation metric. Different from the previous known-unknown evaluation criteria, we propose the concept of "Misunderstand" in LAMA for the first time. Through experiments on 12 PLMs, our context variance prompts and Understand-Confuse-Misunderstand (UCM) metric makes BioLAMA more friendly to large-N-M relations and rare relations. We also conducted a set of control experiments to disentangle "understand" from just "read and copy".