Mitigating Word Bias in Zero-shot Prompt-based Classifiers
This addresses a critical issue for users of zero-shot prompt-based classifiers by improving reliability without labeled data, though it is incremental as it builds on existing prompt-based methods.
The paper tackles the problem of word bias in zero-shot prompt-based classifiers, which causes performance variations with semantically equivalent prompts, by proposing an unsupervised method to reweight class probabilities to have a uniform prior, resulting in large consistent performance gains across NLP tasks.
Prompt-based classifiers are an attractive approach for zero-shot classification. However, the precise choice of the prompt template and label words can largely influence performance, with semantically equivalent settings often showing notable performance difference. This discrepancy can be partly attributed to word biases, where the classifier may be biased towards classes. To address this problem, it is possible to optimise classification thresholds on a labelled data set, however, this mitigates some of the advantages of prompt-based classifiers. This paper instead approaches this problem by examining the expected marginal probabilities of the classes. Here, probabilities are reweighted to have a uniform prior over classes, in an unsupervised fashion. Further, we draw a theoretical connection between the class priors and the language models' word prior, and offer the ability to set a threshold in a zero-resource fashion. We show that matching class priors correlates strongly with the oracle upper bound performance and demonstrate large consistent performance gains for prompt settings over a range of NLP tasks.