Tribrid: Stance Classification with Neural Inconsistency Detection

TL;DR

This paper introduces a neural stance classification model that incorporates negated perspectives and confidence-based filtering to enhance accuracy and reliability in social media analysis.

Contribution

It proposes a novel neural architecture that jointly predicts stance and evaluates inconsistency using negated perspectives, improving classification performance.

Findings

Achieves human-like performance on stance classification tasks.

Using confidence scores effectively filters out doubtful predictions.

Incorporating negated perspectives enhances model robustness.

Abstract

We study the problem of performing automatic stance classification on social media with neural architectures such as BERT. Although these architectures deliver impressive results, their level is not yet comparable to the one of humans and they might produce errors that have a significant impact on the downstream task (e.g., fact-checking). To improve the performance, we present a new neural architecture where the input also includes automatically generated negated perspectives over a given claim. The model is jointly learned to make simultaneously multiple predictions, which can be used either to improve the classification of the original perspective or to filter out doubtful predictions. In the first case, we propose a weakly supervised method for combining the predictions into a final one. In the second case, we show that using the confidence scores to remove doubtful predictions allows our method to achieve human-like performance over the retained information, which is still a sizable part of the original input.