Contrastive Approaches to Climate Change Narrative Classification

Sammanfattning

Climate change is a topic surrounded by many different narratives, either giving ex-planations to the problem and describing solutions, or denying one or more aspects ofit. In recent years, there has been development in natural language processing (NLP)towards computational models for classifying texts by narrative or misinformation claimregarding climate change, following a hierarchical taxonomy of labels. However, trainingmodels using such taxonomies enforce predictions confined to the predefined taxonomy.Ideally, we would like to approach the task in a less supervised way, letting a modelgeneralize to predict outside of the taxonomy it has been trained on, since the climatechange discourse changes through time and between domains.This thesis aims at finding models whose representations are useful across and beyondtaxonomies. As a mean for this, we use the contrastive learning approach as seen inSentence-BERT, and investigate how it can be used for multiclass and multi-label climatechange narrative classification by re-framing the task as a clustering task. We comparecontrastive models to classifiers built on BERT, including cross-dataset and taxonomyperformance. The results show that both classifiers and contrastive models along withclustering can be effectively used for accurate predictions on the task they are trained on.Contrastive models furthermore have the advantage of being able to predict on severallabel granularities and to provide interpretable document similarity scores. However,when evaluating on a dataset with a different taxonomy altogether, we find that bothclassifiers and contrastive models perform poorly.