Will Annotators Disagree? Identifying Subjectivity in Value-Laden Arguments

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Homayounirad, Amir, Liscio, Enrico, Wang, Tong, Jonker, Catholijn M., Siebert, Luciano C.
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908524769968128
author Homayounirad, Amir
Liscio, Enrico
Wang, Tong
Jonker, Catholijn M.
Siebert, Luciano C.
author_facet Homayounirad, Amir
Liscio, Enrico
Wang, Tong
Jonker, Catholijn M.
Siebert, Luciano C.
contents Aggregating multiple annotations into a single ground truth label may hide valuable insights into annotator disagreement, particularly in tasks where subjectivity plays a crucial role. In this work, we explore methods for identifying subjectivity in recognizing the human values that motivate arguments. We evaluate two main approaches: inferring subjectivity through value prediction vs. directly identifying subjectivity. Our experiments show that direct subjectivity identification significantly improves the model performance of flagging subjective arguments. Furthermore, combining contrastive loss with binary cross-entropy loss does not improve performance but reduces the dependency on per-label subjectivity. Our proposed methods can help identify arguments that individuals may interpret differently, fostering a more nuanced annotation process.
format Preprint
id arxiv_https___arxiv_org_abs_2509_06704
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Will Annotators Disagree? Identifying Subjectivity in Value-Laden Arguments
Homayounirad, Amir
Liscio, Enrico
Wang, Tong
Jonker, Catholijn M.
Siebert, Luciano C.
Computation and Language
Aggregating multiple annotations into a single ground truth label may hide valuable insights into annotator disagreement, particularly in tasks where subjectivity plays a crucial role. In this work, we explore methods for identifying subjectivity in recognizing the human values that motivate arguments. We evaluate two main approaches: inferring subjectivity through value prediction vs. directly identifying subjectivity. Our experiments show that direct subjectivity identification significantly improves the model performance of flagging subjective arguments. Furthermore, combining contrastive loss with binary cross-entropy loss does not improve performance but reduces the dependency on per-label subjectivity. Our proposed methods can help identify arguments that individuals may interpret differently, fostering a more nuanced annotation process.
title Will Annotators Disagree? Identifying Subjectivity in Value-Laden Arguments
topic Computation and Language
url https://arxiv.org/abs/2509.06704