Epicurus at SemEval-2023 Task 4: Improving Prediction of Human Values behind Arguments by Leveraging Their Definitions
This work addresses the challenge of predicting subjective human values in arguments, which is incremental as it builds on existing tasks with specific gains.
The paper tackled the problem of identifying human values behind arguments by incorporating definitions of values during model training, resulting in improvements in macro F1 scores of up to 18% compared to baselines.
We describe our experiments for SemEval-2023 Task 4 on the identification of human values behind arguments (ValueEval). Because human values are subjective concepts which require precise definitions, we hypothesize that incorporating the definitions of human values (in the form of annotation instructions and validated survey items) during model training can yield better prediction performance. We explore this idea and show that our proposed models perform better than the challenge organizers' baselines, with improvements in macro F1 scores of up to 18%.