Relative Counterfactual Contrastive Learning for Mitigating Pretrained Stance Bias in Stance Detection

TL;DR

This paper introduces Relative Counterfactual Contrastive Learning (RCCL), a novel approach to reduce pretrained stance bias in stance detection by focusing on relative bias, leading to improved detection performance.

Contribution

It proposes a new structural causal model and a contrastive learning method to effectively mitigate pretrained stance bias in stance detection models.

Findings

RCCL outperforms existing stance detection baselines.

The method effectively reduces pretrained stance bias.

Experimental results demonstrate improved accuracy and fairness.

Abstract

Stance detection classifies stance relations (namely, Favor, Against, or Neither) between comments and targets. Pretrained language models (PLMs) are widely used to mine the stance relation to improve the performance of stance detection through pretrained knowledge. However, PLMs also embed ``bad'' pretrained knowledge concerning stance into the extracted stance relation semantics, resulting in pretrained stance bias. It is not trivial to measure pretrained stance bias due to its weak quantifiability. In this paper, we propose Relative Counterfactual Contrastive Learning (RCCL), in which pretrained stance bias is mitigated as relative stance bias instead of absolute stance bias to overtake the difficulty of measuring bias. Firstly, we present a new structural causal model for characterizing complicated relationships among context, PLMs and stance relations to locate pretrained stance bias. Then, based on masked language model prediction, we present a target-aware relative stance sample generation method for obtaining relative bias. Finally, we use contrastive learning based on counterfactual theory to mitigate pretrained stance bias and preserve context stance relation. Experiments show that the proposed method is superior to stance detection and debiasing baselines.